[go: nahoru, domu]

CN110910549A - Campus personnel safety management system based on deep learning and face recognition features - Google Patents

Campus personnel safety management system based on deep learning and face recognition features Download PDF

Info

Publication number
CN110910549A
CN110910549A CN201911117504.0A CN201911117504A CN110910549A CN 110910549 A CN110910549 A CN 110910549A CN 201911117504 A CN201911117504 A CN 201911117504A CN 110910549 A CN110910549 A CN 110910549A
Authority
CN
China
Prior art keywords
face
card
equipment
attendance
school
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911117504.0A
Other languages
Chinese (zh)
Inventor
张小飞
金良
鲍宇
叶润东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Gaotai Software Technology Co ltd
China University of Mining and Technology CUMT
Original Assignee
Jiangsu Gaotai Software Technology Co ltd
China University of Mining and Technology CUMT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiangsu Gaotai Software Technology Co ltd, China University of Mining and Technology CUMT filed Critical Jiangsu Gaotai Software Technology Co ltd
Priority to CN201911117504.0A priority Critical patent/CN110910549A/en
Publication of CN110910549A publication Critical patent/CN110910549A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161—Detection; Localisation; Normalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168—Feature extraction; Face representation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C1/00—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people
    • G07C1/10—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people together with the recording, indicating or registering of other data, e.g. of signs of identity

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to a campus personnel safety management system based on deep learning and face recognition characteristics, which comprises a background management system and a channel group consisting of a plurality of access control systems, wherein the background management system also comprises teacher and student card information and attendance management, embedded equipment information maintenance and face acquisition App; the channel groups are an obstacle channel group and an obstacle-free channel group, wherein the obstacle channel group is provided with a card swiping device or a video recognition device, the obstacle-free channel group is provided with a video recognition device and a high-definition monitoring camera, when students and teachers swipe campus cards to pass through the obstacle channel group, the card swiping device collects information of the campus cards or the video recognition device collects face images of passing personnel, and the recognition device compares and recognizes campus card identification codes preset in a background system database for the students and teachers according to a recognition method or personal photo information stored in the database, so that whether the students enter a school, leave the school, go in and out of a dormitory through the channel groups or not and whether the students and the teachers normally go out of the dormitory or not are judged.

Description

Campus personnel safety management system based on deep learning and face recognition features
Technical Field
The invention relates to a campus security platform based on the field of biological feature recognition, in particular to a campus personnel security management system based on deep learning and face recognition features by utilizing a face recognition comparison technology and an RFID (radio frequency identification) card recognition technology.
Background
The campus security management platform provides network and information cooperation through a face recognition technology, an access control system, a mobile phone short message, the Internet and the like, and enables a school to master the management platform of school entering and exiting personnel in time.
The platform utilizes face recognition technology. The human face is important information of human beings and is an important basis for distinguishing different people. Therefore, the human face comparison is a more natural and direct comparison method than the technologies such as fingerprints and irises. Face matching is the extraction of specific face feature information from an image or video input and comparison with face feature information registered in a database to obtain the similarity of a matched face and to confirm whether it is the same as a face in the database. The face matching plays an important role in various conditions such as videos, human-computer interfaces, authority control, intelligent monitoring systems and the like.
However, in campus access control, the accuracy and robustness of obtaining the entering person have been the main concern in the industry. The difficulty of utilizing the face recognition technology is caused by lively and active students and concentrated stream of people. Therefore, there is a strong need in the industry for a face-to-face comparison technique that ensures true input, high accuracy and robustness.
The invention provides an embedded face recognition method by combining the existing card service system, promotes the further promotion of an access control system, and has the characteristics of stability and high recognition rate.
Disclosure of Invention
The invention aims to solve the problem that in daily life, students need to go to school and parents cannot communicate with a school in time to know whether the children safely arrive at the school or safely leave the school due to work reasons; the school can not look up and pay attention to each student at any time, so that the problem of the school state of each student can not be known; meanwhile, attendance management can be realized for teachers and students in schools and in the aspect of sleeping, and the system has strong practicability.
The purpose of the invention is realized as follows: a campus personnel safety management system based on deep learning and face recognition features is characterized by comprising a background management system and a plurality of access control system-managed channel groups; the channel groups are divided into barrier channel groups and barrier-free channel groups according to different persons, wherein the barrier channel groups are provided with card swiping equipment or video identification equipment; when students and teachers pass through the channel group by swiping the campus cards, the card swiping equipment collects the information of the campus cards or the video recognition equipment collects the face images of the passing persons and transmits the face images to the recognition equipment through a network, the recognition equipment carries out comparison and recognition on campus card identification codes preset in a background system database by the students and the teachers according to a recognition method or personal photo information stored in the database, after the system inquires the persons in a person database, corresponding in-out time records are made, whether the students enter a school, leave the school, go in and out of a dormitory or not through the channel group or not are recorded, and whether the students and the teachers normally go out of the dormitory or not are recorded; if the person can not be inquired in the database, judging that suspicious persons enter the campus; when an external person without a card enters the campus through the barrier-free channel group, the video recognition device can automatically take a snapshot record, and relevant information can be stored in a background database; in the channel group management, the channel group management is divided into a barrier channel group and an obstacle-free channel group, wherein the barrier channel group is used for card swiping or face recognition, and the obstacle-free channel group is used for card-free access to a campus; card swiping equipment and video identification equipment are arranged in the barrier channel group, video identification equipment is arranged in the barrier-free channel group, and the video identification equipment comprises a high-definition camera, infrared induction equipment and a front-end embedded identification module; the video recognition equipment comprises face image acquisition, face recognition contrast and face deep learning model training in the video recognition process; the human face image acquisition is that when a person enters an infrared sensing range of the camera, the camera starts to shoot and record, and the person in and out images are acquired through frame cutting of the camera; the process comprises the steps of firstly, identifying different frames in a video, uploading images of the frames in which the existence of a mobile figure is found, and identifying faces in the images by a front-end embedded identification module; the process of face recognition comprises face preprocessing, feature extraction, feature comparison and result output; in the image uploaded to the front-end embedded recognition module, the front-end embedded recognition module can utilize a face positioning algorithm to perform image face positioning, feature value extraction is performed after the positioning is successful, the feature value extraction is compared with feature model parameters stored in a memory, and the successful data is judged to be legal personnel by comparison, otherwise, the data is suspicious personnel; the model training comprises the parameters of a human face feature model obtained from a human face database through an experience descriptor and a CNN deep learning algorithm.
Further, the face recognition technology includes the following steps:
step 1) face tracking, acquiring an image difference value, judging the face difference of adjacent frames, entering a new identification process if the difference exceeds a threshold value which is set to be between 0.025 and 0.08, and otherwise discarding the frame;
and 2) judging the image quality, calculating the image definition by using a gradient algorithm, measuring and calculating the brightness by using a gray image mean method, and measuring and calculating the color temperature by using an RGB three-channel or four-channel mean method. Deviation beyond a set threshold; the gradient threshold value needs to be trained, the preset value is 0.2-6, the median value 0.8 of the gray scale map is taken as the lower limit of the brightness value, the median value 1.2 of the gray scale map is taken as the upper limit of the brightness value, the median value 0.7 of the color temperature is taken as the lower limit, and the median value 1.25 of the color temperature is taken as the upper limit;
step 3) extracting detailed face feature data by using a CNN algorithm with more than 64 layers in Tensorflow to form weight output in the CNN algorithm as database feature data;
step 4) face comparison, namely comparing the face feature data with the feature data of each face in a face database to obtain the similarity of the face feature data, wherein the specific method comprises the following steps:
step 4.1) selecting a feature template library of the human face k in the database
Figure BDA0002274470070000031
Step 4.2) to the characteristic template
Figure BDA0002274470070000032
Computing features of an input face
Figure BDA0002274470070000033
And
Figure BDA0002274470070000034
similarity Skji between them;
step 4.3) calculating input face and feature template
Figure BDA0002274470070000035
Degree of similarity of
Figure BDA0002274470070000036
Step 4.4) calculating the similarity between the input face and the face k into
Figure BDA0002274470070000037
Step 4.5) repeating steps 4.1) -4.4), obtaining the similarity between the input face and all K faces in the database, and taking the largest one of the input face and all K faces
Figure BDA0002274470070000038
Obtaining a corresponding human face k'; wherein M is the number of characteristic templates of the person, N is the number of the selected face characteristic points of the person, and i is the face characteristic;
step 5) judging whether a matched face is found; delta is a similarity threshold value set between 0.75 and 0.9 if SmaxIf the value is larger than delta, judging that the input human face is matched with the human face k' in the database;
step 6) judging whether the expression has obvious change; the analysis is performed according to continuous multiframe human face characteristic points, including but not limited to: opening and closing the mouth and eyes, and judging whether the expression of the face is obviously changed;
and 7) outputting the face in the ratio when the facial expression is changed remarkably.
Further, the specific method for extracting detailed face feature data in step 3) is as follows:
step 3.1) according to the accurate human face feature point position obtained by the human face detection and tracking in the step 1), interpolating to obtain the positions of other selected human face feature points;
step 3.2) carrying out normalization processing on the image according to the positions of the two eyes;
step 3.3) calculating to obtain Gabor characteristics of the face characteristic point i
Figure BDA0002274470070000039
The Gabor features of all feature points form face feature data
Figure BDA00022744700700000310
N is the number of the selected face characteristic points; the human face characteristic points are significant characteristic points on human faces, and all 80 Gabor complex coefficients are selected for the characteristics of the human face characteristic points to express complete human face information and completely express differences among different human faces.
Further, step 1) face tracking, wherein the face features selected by the feature points are the features of the commonality of the faces.
Further, the face comparison method also comprises a face registration step, wherein the face characteristic data is stored in a face database, and the specific method comprises the following steps: adding the detailed human face feature data obtained in the step 3) into the human face feature template library of the human
Figure BDA0002274470070000041
M is the number of the characteristic templates of the person and is stored in a database.
Further, the channel group, the card swiping device and the video device are built according to the following steps:
(1) firstly, selecting a place of a school entrance where card swiping equipment or video identification equipment is required to be placed as a channel group, and then establishing equipment information management: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be conveniently interconnected with a platform;
(2) selecting card swiping equipment and video identification equipment in each channel group for installation, setting equipment information of the card swiping equipment, and managing the card swiping equipment information: aiming at a management system for uploading and downloading of a card swiping device, an RS485 network cable is used for connecting a background server, identification is carried out through an IP address, and an ftp server started by an embedded Linux system of a card machine is used for managing personnel RFID card serial number information in the card machine;
the face recognition equipment comprises a camera and a front-end embedded recognition module; the camera is connected with the embedded module through an RS485 or USB line, and the front-end embedded identification module is connected with the background server through an RS485 network cable; the identification parameter setting mainly refers to the updating of face training result parameters and the updating of a face feature database of legal personnel by connecting the front end with an embedded Linux system through an FTP (file transfer protocol);
(3) the reporting state is as follows: the card swiping machine is set to report the state of the machine on time, and the uploaded information comprises the serial number and ip address of card swiping equipment, the starting time and the shutdown time of the last operation of the equipment, the restarting time of the card swiping machine, the data modification and deletion time stored at the front end of the card swiping machine, and the personnel card access record in the operation period;
(4) the teacher and the parents upload personal photos of the students and the school entering personnel through 'simplified to people communication app', and the personal photos are stored in a database and used for face recognition;
(5) using a background server to learn the face characteristics according to information such as photos, fingerprints and the like, acquiring learning data parameters, downloading the data parameters by an administrator, and storing the data parameters into a memory of a card swiping device or a front-end embedded identification module; the card machine data is directly downloaded to the front-end card swiping machine without learning;
(6) when the school needs to add more channel groups and card swiping equipment, the steps (1) - (5) are repeated.
Furthermore, the background management system comprises a teacher and student card information and attendance management module, an embedded equipment information maintenance module and a face information acquisition app module.
Furthermore, the teacher and student information and attendance management module comprises card affair management, attendance management, statistical inquiry and personnel access. The card affair management manages campus cards, the campus cards are divided into student cards, teacher cards, physical cards and 2.4G active cards, and the cards correspondingly contain student information, teacher information, physical card information, 2.4G active card information and face recognition information; the school administrator binds the card number in advance and then imports the card number; a teacher downloads a template according to the class, information data of teachers and students are filled in the template or card numbers are collected through a card sender, personnel correspond to the card numbers one by one, and the information data of the personnel are synchronized into the card; the card information comprises student names, guardian telephones, school numbers, classes, card information, card validity periods, use records, use time and face data; if the 2.4G active card is used, the radio frequency of 2.4GHZ is adopted, and when the electric quantity is lower than a certain value, the card swiping equipment gives out low-electric-quantity alarm when the wireless signal strength is insufficient.
The attendance management comprises student attendance, teacher attendance, dormitory attendance, in-school state monitoring and attendance setting. When students and teachers pass through the channel group by swiping the campus cards, the card swiping equipment collects the information of the campus cards or the video recognition equipment collects the face images of the passing persons and transmits the face images to the recognition equipment through a network, the recognition equipment compares and recognizes the campus card identification codes preset in a background system database by the students and the teachers according to the recognition method or compares and recognizes the personal photo information stored in the database, and after the system inquires the persons in the person database, the system makes corresponding entry and exit time records and monthly attendance statistical records so as to record whether the students enter a school, leave the school, go in and out of a dormitory through the channel group or not and whether the students and the teachers normally go out of the dormitory or not; in the school state query, if students go in and out of a school or a dormitory by swiping cards, the states of the students in the school and the dormitory can be dynamically updated and changed, and the states of the students in the school and the dormitory can be conveniently and visually observed; in the attendance setting, the attendance time of school attendance equipment is included, five time periods including school entrance, school exit, late arrival, early exit and freedom are included, and a school administrator can set the attendance time period according to the specific requirements of a school; the attendance statistical short message can be set with the sending time of the school attendance statistical short message, whether the statistical short message is sent to the teacher every Saturday or not and the attendance result.
The statistical query comprises student attendance conditions, staff attendance conditions, dormitory attendance conditions and kindergarten refund report forms, and the student attendance conditions can be optionally queried, or within a week and a month, the student attendance conditions. The attendance checking and counting short message function is also included in the attendance checking and counting query of students, and the attendance checking and counting short message function can be set in the attendance checking and counting short message, whether the statistics short message is sent to teachers or parents on saturday or not and the attendance checking result; the attendance condition of the teaching staff comprises statistics of attendance, absence and leave requests of the teaching staff within one month; the dormitory attendance condition comprises statistics of the state of students in dormitories, and students leave the dormitories or return home to the dormitories within the sleeping time and are classified as not attendance; the kindergarten refund report contains the use and refund condition of each expense in daily life, and the parents and the users can conveniently check the refund report when needed.
The personnel access comprises card-free personnel access management and special personnel account management, and the card-free personnel access management comprises card-free personnel registration and card-free personnel snapshot record. When the students or teachers do not carry the campus card, the students or teachers can enter or leave the campus after registering information through related personnel at the barrier-free passage group, and manual sign-in and sign-out records are recorded into attendance statistics; when an external card-free person enters the campus, the video recognition device can automatically take a snapshot record, and if the card-free person carries out check-in registration, registration information can be stored in a background database; the special personnel account management comprises grade level master-and-master account management and dormitory manager account management, and the grade master-and-master account management and the dormitory manager can log in different management interfaces respectively.
Further, the embedded information maintenance module comprises equipment information management, a place where card swiping equipment or video identification equipment needs to be placed at an entrance and an exit of a school is selected as a channel group, and then equipment information management is established: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be connected with a platform. The card swiping equipment comprises an RFID card swiping machine, a 2.4GHZ identifier and fingerprint identification equipment, wherein the RFID card swiping machine is used for identifying student cards, teacher cards and physical cards, and the 2.4GHZ identifier is mainly used for identifying 2.4GHZ active cards; the fingerprint identification device is used for fingerprint identification; the video identification device comprises a high-definition camera, an infrared sensing device and a front-end embedded identification module; the minimum requirement of the high-definition camera is 200 ten thousand pixels, and the frame rate is 25 fps; the front-end embedded identification chip comprises an ARM main control chip, a GPU chip supporting deep learning and an external monitoring camera interface, wherein the external monitoring camera interface comprises but is not limited to an MIPI interface, a DVP interface and a CSI interface.
Further, the face information acquisition app enables parents to register and log in personal photos, firstly downloads ' brief to people ' through app ' to select to register and log in, uploads the personal photos after entering a personal center, stores the personal photos in a database for face recognition, can click a video attendance check to select to inquire records of children entering and exiting a school after uploading is finished, and can view the identification photos of the children entering and exiting the school at icons behind the corresponding records.
The teacher binds the card and uploads the personal photo, firstly downloads the ' simplified to people ' app ' to select registration login, registers on the app and uploads, and after the registration login and the uploading, the teacher can upload and update the student identification in the designated range. In addition, the teacher can check whether the students safely leave the school or not by clicking the 'safety leave the school' function, and can check the identification photos of the students or the teachers entering and leaving the school by clicking the video attendance.
Whether the school authorizes the teacher to manage the student identification photo or not needs to use the identity of a school administrator to log in background management software, authority setting in face identification photo information is selected, authority of the whole teacher or the designated teacher to manage the student identification photo is granted, and therefore the student photo can be uploaded to a background management part.
Has the advantages that: 1) as the school adopts the software, in the aspect of school, a school administrator can easily master the state of the students in the school, whether the students safely enter or leave the school or not, and the accommodation school can conveniently know whether the students are in a dormitory or not; parents can conveniently inquire the state of the child in the school, the safety problem of the child cannot be neglected due to working reasons, and the parents can further confirm whether the child safely enters or leaves the school through video attendance; besides paying attention to students in class, the teacher can know whether the students leave the school safely or not after leaving class, and can pay attention to the safety state of the students in the whole course. 2) In the aspect of attendance of teachers and students, teachers can know attendance statistics of students for one week on saturday days, and therefore the teachers can conveniently communicate with parents to inform children of attendance conditions; the school can easily master the attendance conditions of students and teachers, whether the students and the teachers are absent and the number of days of absence, and the contents are clear at a glance. 3) The channel group in the invention can be optionally added, modified or deleted according to the school requirements, the identification equipment in the channel group is not invariable, one or more functions can be selected according to the requirements, and only card swiping or face identification is carried out, so that the expenditure can be reduced. 4) The invention has lower cost and simple and convenient system construction, can be used for some sudden scenes and quickly and accurately acquire the required information. 5) The invention has strong universality, applicability and expandability, can be applied to different environments and scenes especially in the aspect of safety, and has strong practicability. 6) The human face comparison method eliminates the influence of human face expression and posture, judges the authenticity of the human face in comparison, and ensures that the tracking and comparison accuracy, precision and robustness are higher.
Drawings
Fig. 1 is a system structure diagram of a campus personnel security management system based on deep learning and face recognition features.
Fig. 2 is an equipment information diagram of a channel group composed of a plurality of access control systems.
Fig. 3 is a detailed structure diagram of the background management system.
Fig. 4 is a timing diagram of data reporting of the attendance checking device.
Fig. 5 is a flowchart of a face comparison method.
Fig. 6 is a flow chart of face recognition data.
FIG. 7 is a flow chart of card swipe data processing.
Fig. 8 is a flow chart of the school opening student card.
Fig. 9 is a diagram illustrating a case where the operation and maintenance staff opens the video service.
Fig. 10 is a flow chart of student's certificate card swiping picture identification.
FIG. 11 is an illustration of school administrator student identity card business.
Detailed Description
Example 1: fig. 1 is a system structure diagram, a school needs to be provided with a plurality of channel groups within a maintenance range, each channel group is equipped with a corresponding card swiping device and a corresponding video identification device, the devices are connected with a background management system, and the background management system provides a data access function for the devices in the channel groups through a database through network connection. The school administrator uploads the card affair information and the personal identification pictures of all teachers and students to the background system database, when students and teachers pass through the channel group card swiping equipment, the card swiping equipment collects the campus card information or the video identification equipment collects the face images of the passing personnel, the identification equipment compares and identifies the campus card identification codes preset in the background system database by the students and the teachers or compares and identifies the personal photo information stored in the database, corresponding records are stored in the database, and the records are displayed in corresponding modules of the background management system, so that whether the students safely enter schools, leave schools and go in and out dormitories through the channel group or not is known, and whether the students and the teachers normally go out of the dormitories or not is known.
Fig. 2 is an equipment information diagram of a plurality of access control system managed access groups, in the access group management, configuration information of each access group is different and is divided into an obstacle access group and a non-obstacle access group, the obstacle access group can be used for students or teachers to swipe cards to pass through, and the non-obstacle access group is mainly used for non-card persons to enter a campus. A card swiping device and a video identification device are arranged in the barrier passage group, the card swiping device comprises an RFID card swiping machine, a 2.4GHZ identifier and a fingerprint identification device, the RFID card swiping machine is used for identifying a student card, a teacher card and a physical card, and the 2.4GHZ identifier is mainly used for identifying a 2.4GHZ active card; the fingerprint identification device is used for fingerprint identification; the video identification device comprises a high-definition camera, an infrared sensing device and a front-end embedded identification module; the high-definition camera is used for an obstacle-free channel group, the minimum requirement of pixels is 200 ten thousand pixels, and the frame rate is 25 fps; the front-end embedded recognition chip is used for the barrier channel group and comprises an ARM main control chip, a GPU chip supporting deep learning and an external monitoring camera interface, wherein the external interface comprises but is not limited to an MIPI interface, a DVP interface and a CSI interface
The construction and use steps of the channel group and the attendance card swiping equipment are as follows:
(1) firstly, selecting a place of a school entrance where card swiping equipment or video identification equipment is required to be placed as a channel group, and then establishing equipment information management: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be conveniently interconnected with a platform;
(2) selecting card swiping equipment and video identification equipment in each channel group for installation, setting equipment information of the card swiping equipment, and managing the card swiping equipment information: aiming at a management system for uploading and downloading card swiping equipment, an RS485 network cable is used for connecting a background server, identification is carried out through an IP address, and personnel RFID card serial number information in a card machine is managed by an FTP server started by an embedded Linux system of the card machine;
the face recognition equipment comprises a camera and a front-end embedded recognition module; the camera is connected with the embedded module through an RS485 or USB line, and the front-end embedded identification module is connected with the background server through an RS485 network cable; the identification parameter setting mainly refers to the updating of face training result parameters and the updating of a face feature database of legal personnel by connecting the front end with an embedded Linux system through an FTP (file transfer protocol);
(3) the reporting state is as follows: the card swiping machine is set to report the state of the machine on time, and the uploaded information comprises the serial number and the IP address of the card swiping equipment, the starting time and the shutdown time of the last operation of the equipment, the restarting time of the card swiping machine, the data modification and deletion time stored at the front end of the card swiping machine, and the entrance and exit records of the personnel card in the operation period;
(4) the teacher and the parents upload personal photos of the students and the school entering personnel through 'simplified to people communication app', and the personal photos are stored in a database and used for face recognition;
(5) using a background server to learn the face characteristics according to information such as photos, fingerprints and the like, acquiring learning data parameters, downloading the data parameters by an administrator, and storing the data parameters into a memory of a card swiping device or a front-end embedded identification module; the card machine data is directly downloaded to the front-end card swiping machine without learning;
(6) when the school needs to add more channel groups and card swiping equipment, the steps (1) - (5) are repeated.
Fig. 3 is a detailed structure diagram of a background management system, which comprises a teacher and student card information and attendance management module, an embedded information maintenance module and a face information acquisition APP module in the background management system.
The teacher and student information and attendance management module also comprises card affair management, attendance management, statistical inquiry and personnel access. The card affair management manages campus cards, the campus cards are divided into student cards, teacher cards, physical cards and 2.4G active cards, and the cards correspondingly contain student information, teacher information, physical card information, 2.4G active card information and face recognition information; the school administrator binds the card number in advance and then imports the card number; a teacher downloads a template according to the class, information data of teachers and students are filled in the template or card numbers are collected through a card sender, personnel correspond to the card numbers one by one, and the information data of the personnel are synchronized into the card; the card information comprises student names, guardian telephones, school numbers, classes, card information, card validity periods, use records, use time and face data; if the 2.4G active card is used, the radio frequency of 2.4GHZ is adopted, and when the electric quantity is lower than a certain value, the card swiping equipment gives out low-electric-quantity alarm when the wireless signal strength is insufficient.
The attendance management comprises student attendance, teacher attendance, dormitory attendance, in-school state monitoring and attendance setting. When students and teachers pass through the channel group by swiping the campus cards, the card swiping equipment collects the information of the campus cards or the video recognition equipment collects the face images of the passing persons and transmits the face images to the recognition equipment through a network, the recognition equipment compares and recognizes the campus card identification codes preset in a background system database by the students and the teachers according to the recognition method or compares and recognizes the personal photo information stored in the database, and after the system inquires the persons in the person database, the system makes corresponding entry and exit time records and monthly attendance statistical records so as to record whether the students enter a school, leave the school, go in and out of a dormitory through the channel group or not and whether the students and the teachers normally go out of the dormitory or not; in the school state query, if students go in and out of a school or a dormitory by swiping cards, the states of the students in the school and the dormitory can be dynamically updated and changed, and the states of the students in the school and the dormitory can be conveniently and visually observed; in the attendance setting, the attendance time of school attendance equipment is included, five time periods including school entrance, school exit, late arrival, early exit and freedom are included, and a school administrator can set the attendance time period according to the specific requirements of a school; the attendance statistical short message can be set with the sending time of the school attendance statistical short message, whether the statistical short message is sent to the teacher every Saturday or not and the attendance result.
The statistical query comprises student attendance conditions, staff attendance conditions, dormitory attendance conditions and kindergarten refund report forms, and the student attendance conditions can be optionally queried, or within a week and a month, the student attendance conditions. The attendance checking and counting short message function is also included in the attendance checking and counting query of students, and the attendance checking and counting short message function can be set in the attendance checking and counting short message, whether the statistics short message is sent to teachers or parents on saturday or not and the attendance checking result; the attendance condition of the teaching staff comprises statistics of attendance, absence and leave requests of the teaching staff within one month; the dormitory attendance condition comprises statistics of the state of students in dormitories, and students leave the dormitories or return home to the dormitories within the sleeping time and are classified as not attendance; the kindergarten refund report contains the use and refund condition of each expense in daily life, and the parents and the users can conveniently check the refund report when needed.
The personnel access comprises card-free personnel access management and special personnel account management, and the card-free personnel access management comprises card-free personnel registration and card-free personnel snapshot record. When the students or teachers do not carry the campus card, the students or teachers can enter or leave the campus after registering information through related personnel at the barrier-free passage group, and manual sign-in and sign-out records are recorded into attendance statistics; when an external card-free person enters the campus, the video recognition device can automatically take a snapshot record, and if the card-free person carries out check-in registration, registration information can be stored in a background database; the special personnel account management comprises grade level master-and-master account management and dormitory manager account management, and the grade master-and-master account management and the dormitory manager can log in different management interfaces respectively.
In the embedded information maintenance module containing equipment information management, firstly, selecting a place of a school entrance where card swiping equipment or video identification equipment is required to be placed as a channel group, and then, establishing equipment information management: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be connected with a platform. The card swiping equipment comprises an RFID card swiping machine, a 2.4GHZ identifier and fingerprint identification equipment, wherein the RFID card swiping machine is used for identifying student cards, teacher cards and physical cards, and the 2.4GHZ identifier is mainly used for identifying 2.4GHZ active cards; the fingerprint identification device is used for fingerprint identification; the video identification device comprises a high-definition camera, an infrared sensing device and a front-end embedded identification module; the minimum requirement of the high-definition camera is 200 ten thousand pixels, and the frame rate is 25 fps; the front-end embedded identification chip comprises an ARM main control chip, a GPU chip supporting deep learning and an external monitoring camera interface, wherein the external monitoring camera interface comprises but is not limited to an MIPI interface, a DVP interface and a CSI interface.
In the face information acquisition app, parents register and log in and upload personal photos, firstly, a ' brief to people ' communication app ' selection is downloaded to perform registration and logging in, the personal photos are uploaded after entering a personal center and stored in a database for face recognition, after uploading is completed, a user can click a video attendance check selection to inquire the records of children entering and exiting a school, and the user can view the identification photos of the children entering and exiting the school at icons behind the corresponding records.
The teacher binds the card and uploads the personal photo, firstly downloads the ' simplified to people ' app ' to select registration login, registers on the app and uploads, and after the registration login and the uploading, the teacher can upload and update the student identification in the designated range. In addition, the teacher can check whether the students safely leave the school or not by clicking the 'safety leave the school' function, and can check the identification photos of the students or the teachers entering and leaving the school by clicking the video attendance.
Whether the school authorizes teachers to manage student identification photos or not needs to use school administrator identities to log in background management software, authority setting in face identification photo management is selected, authority of all teachers or appointed teachers to manage the identification photos is granted, and therefore the student photos can be uploaded to a background management part.
Fig. 4 is a timing diagram of data reporting of the attendance checking device. The attendance checking equipment is divided into three states: when a student or a teacher swipes the card through the barrier channel group, the card swiping record enters a queue for waiting for service, and the card swiping result is matched with the database and returned after the card swiping record is calculated by the attendance checking calculation component and the record is stored; when a student or a teacher or an off-school person does not carry a campus card channel barrier-free channel group, the video identification equipment identifies the student or the teacher or the off-school person, stores a card-free access record, enters a queue for waiting for service, calculates an attendance result through the calculation component, matches the attendance result with the database and stores the attendance record; the attendance checking equipment reports the equipment state on time and stores the equipment state in a database.
Fig. 5 is a flow chart of face recognition data. The method comprises the steps of obtaining face image information through hardware video equipment, carrying out model training and face recognition comparison, preprocessing data, extracting features, comparing the features and outputting results.
The video recognition process mainly comprises face image acquisition, face recognition comparison and face deep learning model training.
The human face image acquisition is that when a person enters an infrared sensing range of the camera, the camera starts shooting and recording, and the person in and out images are acquired through frame cutting of the camera. The process comprises the steps of firstly identifying different frames in a video, uploading images of the frames in which the existence of the mobile character is found, and identifying the face in the images by a front-end embedded identification module.
The face recognition process mainly comprises face preprocessing, feature extraction, feature comparison and result output. In the image uploaded to the front-end embedded recognition module, the front-end embedded recognition module can utilize a face positioning algorithm to perform image face positioning, feature value extraction is performed after the positioning is successful, the feature value extraction is compared with feature model parameters stored in a memory, and the successful data is judged to be legal personnel through comparison, otherwise, the data is suspicious personnel.
The model training mainly comprises the face characteristic model parameters obtained from a face database through an experience descriptor and a CNN deep learning algorithm.
Fig. 6 is a flowchart of a face comparison method. And extracting characteristic points from the obtained face data, and comparing the characteristic points with data stored in a database. And detecting the quality of the video image and identifying the change of the facial expression of the human face.
The image recognition relates to a face recognition technology, and comprises the following steps:
step 601, tracking a human face, acquiring an image difference value, judging the human face difference of adjacent frames, if the difference exceeds a threshold value, setting the threshold value to be between 0.025 and 0.08, entering a new identification process, and if not, discarding the frame;
step 602, judging image quality, calculating image definition by using a gradient algorithm, measuring and calculating brightness by using a gray image mean method, and measuring and calculating color temperature by using an RGB three-channel or four-channel mean method. Deviation beyond a set threshold; the gradient threshold value needs training, the preset value is 0.2-6, the median value 0.8 of the brightness value is taken as the lower limit, the median value 1.2 of the brightness value is taken as the upper limit, the median value 0.7 of the color temperature is taken as the lower limit, and the median value 1.25 of the color temperature is taken as the upper limit.
Step 603, extracting detailed human face feature data by using a CNN algorithm of more than 64 layers in the tenserflow to form weight output in the CNN algorithm as database feature data;
step 605, comparing the human faces, namely comparing the human face feature data with the feature data of each human face in the human face database to obtain the similarity of the human face feature data and each human face in the human face database; the specific method comprises the following steps:
selecting a feature template library of a face k in a database
Figure BDA0002274470070000121
Template for characteristics
Figure BDA0002274470070000122
Computing features of an input face
Figure BDA0002274470070000123
And
Figure BDA0002274470070000124
similarity Skji between them;
calculating input face and feature template
Figure BDA0002274470070000125
Degree of similarity of
Figure BDA0002274470070000126
Calculating the similarity between the input face and the face k as
Figure BDA0002274470070000127
Repeating the steps (1) to (4) to obtain the similarity between the input human face and all K human faces in the database, and taking the largest one of the input human faces
Figure BDA0002274470070000128
Obtaining a corresponding human face k';
wherein M is the number of characteristic templates of the person, N is the number of the selected face characteristic points of the person, and i is the face characteristic.
Step 607, judging whether a matched face is found; delta is a similarity threshold value set between 0.75 and 0.9 if SmaxIf delta is greater, the input human face and number are judgedMatching the human faces k' in the database;
step 608, judging whether the expression has significant changes; the analysis is performed according to continuous multiframe human face characteristic points, including but not limited to: opening and closing the mouth and eyes, and judging whether the expression of the face is obviously changed;
when the facial expression has a significant change, step 609 is executed to output the face in the ratio.
The specific method for extracting the detailed face feature data in step 603 is as follows:
interpolating to obtain the positions of other selected human face characteristic points according to the accurate human face characteristic point positions obtained by the human face detection and tracking in the step 601;
normalizing the image according to the positions of the two eyes;
calculating to obtain Gabor characteristics of the face characteristic point i
Figure BDA0002274470070000131
The Gabor features of all feature points form face feature data
Figure BDA0002274470070000132
And N is the number of the selected face characteristic points.
The human face characteristic points are significant characteristic points on human faces, and all 80 Gabor complex coefficients are selected for the characteristics of the human face characteristic points to express complete human face information and completely express differences among different human faces.
In this step 601, face tracking is performed, and the face features selected by the feature points are features of commonality of faces.
Further, the face comparison method further comprises a step 604 of face registration; storing the face characteristic data to a face database; the specific method comprises the following steps:
adding the detailed face feature data obtained in the step 603 into the face feature template library of the person
Figure BDA0002274470070000133
Figure BDA0002274470070000134
M is the number of the characteristic templates of the person and is stored in a database.
FIG. 7 is a flow chart of card swipe data processing. When a user swipes a card on the card swiping equipment, the card swiping equipment reports a card swiping record, the card swiping record is received by a platform or equipment middleware and then enters a service queue, the data legality is judged after the data is listed out, if the data is stored in a database, whether a short message is sent or not is selected, if the data is sent, the data enters a short message sending queue, and if the data is not sent, an attendance checking result is stored.
Fig. 8 is a flow chart of student identity card service provisioning. The school applies for student identity card service, and the system collects field conditions, selects a school to be maintained and detects whether the school has opened student identity card service. After the service is opened, the system can correspondingly set the channel group, the attendance equipment, the card service, the attendance mode and the like, and issues the card to the user to perform corresponding service test to ensure that the service has no errors.
Fig. 9 is a flow chart of student's certificate card swiping picture identification. When a student holds a card and swipes the card at the card swiping equipment of the obstacle passage group, the system caches the card swiping data, inquires the cached personnel data and pictures and displays the data and the pictures in the card swiping equipment.
Fig. 10 is a diagram of a parent-student identity card business. After applying for opening the student identity card service by the school, the operation and maintenance personnel open the student identity card service for the parents. The business comprises services such as student card swiping identification photo management, student in-and-out condition query, student attendance condition statistics query and the like.
FIG. 11 is an illustration of a school administrator student's certificate. The school administrator mainly has four management modules of card affair management, attendance management, statistical inquiry and personnel access
The face recognition technology can construct a multi-layer structure face model to adapt to the change of face expression, can also construct face shape models at different angles to adapt to the change of face angles, and can recognize the face appearance of students without errors.

Claims (10)

1. A campus personnel safety management system based on deep learning and face recognition features is characterized by comprising a background management system and a channel group consisting of a plurality of access control systems; the channel group is divided into an obstacle channel group and a barrier-free channel group according to different personnel passing modes, wherein the obstacle channel group is provided with card swiping equipment or video identification equipment; when students and teachers pass through the barrier channel group by swiping the campus cards, acquiring information of the campus cards by swiping card equipment, or acquiring face images of passing personnel by video recognition equipment, and transmitting the face images to the recognition equipment through a network; after the system inquires the personnel in the personnel database, recording corresponding entry and exit time records, and distinguishing whether students enter a school, leave the school and go in and out of a dormitory through a channel group and whether students and teachers normally work or not; when the person can not be inquired in the database, judging that suspicious persons enter the campus; when an external person without a card enters the campus through the barrier-free channel group, the video recognition device can automatically take a snapshot record, and relevant information can be stored in the background database and is notified to relevant persons; in the channel group management, the barrier channel group supports card swiping or a face recognition method to pass through, and the barrier-free channel group is used for card-free access to a campus; card swiping equipment and video identification equipment are arranged in the barrier channel group, video identification equipment is arranged in the barrier-free channel group, and the video identification equipment comprises a high-definition camera, infrared induction equipment and a front-end embedded identification module; the video recognition method comprises the steps of face image acquisition, face recognition contrast and face deep learning model training in the video recognition process;
the human face image acquisition is that when a person enters an infrared sensing range of the camera, the camera starts to shoot and record, and the person in and out images are acquired through frame cutting of the camera; the process comprises the steps of firstly, identifying different frames in a video, uploading images of the frames in which the existence of a mobile figure is found, and identifying faces in the images by a front-end embedded identification module;
the process of face recognition comprises face preprocessing, feature extraction, feature comparison and result output; in the image uploaded to the front-end embedded recognition module, the front-end embedded recognition module can utilize a face positioning algorithm to perform image face positioning, feature value extraction is performed after the positioning is successful, the feature value extraction is compared with feature model parameters stored in a memory, and the successful data is judged to be legal personnel by comparison, otherwise, the data is suspicious personnel;
the model training comprises the parameters of a human face feature model obtained from a human face database through experience description factors and a CNN deep learning algorithm.
2. The campus personnel safety management system based on deep learning and face recognition features as claimed in claim 1, characterized in that: the technology for recognizing the human face is characterized by comprising the following steps of:
step 1) face tracking, acquiring an image difference value, judging the face difference of adjacent frames, entering a new identification process if the difference exceeds a threshold value which is set to be between 0.025 and 0.08, and otherwise discarding the frame;
step 2) judging the image quality, calculating the image definition by using a gradient algorithm, measuring and calculating the brightness by using a gray image mean method, and measuring and calculating the color temperature by using an RGB three-channel or four-channel mean method; deviation beyond a set threshold; the gradient threshold value needs to be trained, the preset value is 0.2-6, the median value 0.8 of the gray scale map is taken as the lower limit of the brightness value, the median value 1.2 of the gray scale map is taken as the upper limit of the brightness value, the median value 0.7 of the color temperature is taken as the lower limit, and the median value 1.25 of the color temperature is taken as the upper limit;
step 3) extracting detailed face feature data by using a CNN algorithm with more than 64 layers in Tensorflow to form weight output in the CNN algorithm as database feature data;
step 4) face comparison, namely comparing the face feature data with the feature data of each face in a face database to obtain the similarity of the face feature data, wherein the specific method comprises the following steps:
step 4.1) selecting a feature template library of the human face k in the database
Figure FDA0002274470060000021
Step 4.2) to the characteristic template
Figure FDA0002274470060000022
Computing features of an input face
Figure FDA0002274470060000023
And
Figure FDA0002274470060000024
similarity between them Skji；
Step 4.3) calculating input face and feature template
Figure FDA0002274470060000025
The similarity of (2);
Figure FDA0002274470060000026
step 4.4) calculating the similarity between the input face and the face k into
Figure FDA0002274470060000027
Step 4.5) repeating steps 4.1) -4.4), obtaining the similarity between the input face and all K faces in the database, and taking the largest one of the input face and all K faces
Figure FDA0002274470060000028
Obtaining a corresponding human face k'; wherein M is the number of characteristic templates of the person, N is the number of the selected face characteristic points of the person, and i is the face characteristic;
step 5) judging whether a matched face is found; delta is a similarity threshold value set between 0.75 and 0.9 if SmaxIf the value is larger than delta, judging that the input human face is matched with the human face k' in the database;
step 6) judging whether the expression has obvious change; the analysis is performed according to continuous multiframe human face characteristic points, including but not limited to: opening and closing the mouth and eyes, and judging whether the expression of the face is obviously changed;
and 7) outputting the face in the ratio when the facial expression is changed remarkably.
3. The campus personnel safety management system based on deep learning and face recognition features as claimed in claim 2, wherein the specific method for extracting detailed face feature data in step 3) is as follows:
step 3.1) according to the accurate human face feature point position obtained by the human face detection and tracking in the step 1), interpolating to obtain the positions of other selected human face feature points;
step 3.2) carrying out normalization processing on the image according to the positions of the two eyes;
step 3.3) calculating to obtain Gabor characteristics of the face characteristic point i
Figure FDA0002274470060000031
The Gabor features of all feature points form face feature data
Figure FDA0002274470060000032
N is the number of the selected face characteristic points; the human face characteristic points are significant characteristic points on human faces, and all 80 Gabor complex coefficients are selected for the characteristics of the human face characteristic points to express complete human face information and completely express differences among different human faces.
4. The campus personnel security management system based on deep learning and face recognition features as claimed in claim 1, wherein step 1) face tracking, the face features selected by the feature points are obtained as the features of commonality of faces.
5. The campus personnel safety management system based on deep learning and face recognition features as claimed in claim 3, wherein the face comparison method further comprises a face registration step, the face feature data is stored in a face database, and the specific method is as follows: adding the detailed human face feature data obtained in the step 3) into the human face feature template library of the human
Figure FDA0002274470060000033
M is the number of the characteristic templates of the person and is stored in a database.
6. The campus personnel safety management system based on deep learning and face recognition features as claimed in claim 1, wherein the channel group, the card swiping device and the video device are constructed by the following steps:
(1) firstly, selecting a place of a school entrance where card swiping equipment or video identification equipment is required to be placed as a channel group, and then establishing equipment information management: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be conveniently interconnected with a platform;
(2) selecting card swiping equipment and video identification equipment in each channel group for installation, setting equipment information of the card swiping equipment, and managing the card swiping equipment information: aiming at a management system for uploading and downloading card swiping equipment, an RS485 network cable is used for connecting a background server, identification is carried out through an IP address, and personnel RFID card serial number information in a card machine is managed by an FTP server started by an embedded Linux system of the card machine;
the face recognition equipment comprises a camera and a front-end embedded recognition module; the camera is connected with the embedded module through an RS485 or USB line, and the front-end embedded identification module is connected with the background server through an RS485 network cable; the identification parameter setting mainly refers to the updating of face training result parameters and the updating of a face feature database of legal personnel by connecting the front end with an embedded Linux system through an FTP (file transfer protocol);
(3) the reporting state is as follows: the card swiping machine is set to report the state of the machine on time, and the uploaded information comprises the serial number and the IP address of the card swiping equipment, the starting time and the shutdown time of the last operation of the equipment, the restarting time of the card swiping machine, the data modification and deletion time stored at the front end of the card swiping machine, and the entrance and exit records of the personnel card in the operation period;
(4) the teacher and the parents upload personal photos of the students and the school entering personnel through 'simplified to people communication app', and the personal photos are stored in a database and used for face recognition;
(5) using a background server to learn the face characteristics according to information such as photos, fingerprints and the like, acquiring learning data parameters, downloading the data parameters by an administrator, and storing the data parameters into a memory of a card swiping device or a front-end embedded identification module; the card machine data is directly downloaded to the front-end card swiping machine without learning;
(6) when the school needs to add more channel groups and card swiping equipment, the steps (1) - (5) are repeated.
7. The campus personnel safety management system based on deep learning and face recognition features as claimed in claim 1, wherein the background management system comprises a teacher and student card information and attendance management module, an embedded equipment information maintenance module, and a face information acquisition App module.
8. The campus personnel security management system based on deep learning and face recognition features as claimed in claim 1, wherein the teacher and student card affair information and attendance management module further comprises card affair management, attendance management, statistical query and personnel admission. The card affair management manages campus cards, the campus cards are divided into student cards, teacher cards, physical cards and 2.4G active cards, and the cards correspondingly contain student information, teacher information, physical card information, 2.4G active card information and face recognition information; the school administrator binds the card number in advance and then imports the card number; a teacher downloads a template according to the class, information data of teachers and students are filled in the template or card numbers are collected through a card sender, personnel correspond to the card numbers one by one, and the information data of the personnel are synchronized into the card; the card information comprises student names, guardian telephones, school numbers, classes, card information, card validity periods, use records, use time and face data; if the 2.4G active card is used, the radio frequency of 2.4GHZ is adopted, and when the electric quantity is lower than a certain value, the card swiping equipment gives out low-electric-quantity alarm when the wireless signal strength is insufficient.
The attendance management comprises student attendance, teacher attendance, dormitory attendance, in-school state monitoring and attendance setting. When students and teachers pass through the channel group by swiping the campus cards, the card swiping equipment collects the information of the campus cards or the video recognition equipment collects the face images of the passing persons and transmits the face images to the recognition equipment through a network, the recognition equipment compares and recognizes the campus card identification codes preset in a background system database by the students and the teachers according to the recognition method or compares and recognizes the personal photo information stored in the database, and after the system inquires the persons in the person database, the system makes corresponding entry and exit time records and monthly attendance statistical records so as to record whether the students enter a school, leave the school, go in and out of a dormitory through the channel group or not and whether the students and the teachers normally go out of the dormitory or not; in the school state query, if students go in and out of a school or a dormitory by swiping cards, the states of the students in the school and the dormitory can be dynamically updated and changed, and the states of the students in the school and the dormitory can be conveniently and visually observed; in the attendance setting, the attendance time of school attendance equipment is included, five time periods including school entrance, school exit, late arrival, early exit and freedom are included, and a school administrator can set the attendance time period according to the specific requirements of a school; the attendance statistical short message can be set with the sending time of the school attendance statistical short message, whether the statistical short message is sent to the teacher every Saturday or not and the attendance result.
The statistical query comprises student attendance conditions, staff attendance conditions, dormitory attendance conditions and kindergarten refund report forms, and the student attendance conditions can be optionally queried, or within a week and a month, the student attendance conditions. The attendance checking and counting short message function is also included in the attendance checking and counting query of students, and the attendance checking and counting short message function can be set in the attendance checking and counting short message, whether the statistics short message is sent to teachers or parents on saturday or not and the attendance checking result; the attendance condition of the teaching staff comprises statistics of attendance, absence and leave requests of the teaching staff within one month; the dormitory attendance condition comprises statistics of the state of students in dormitories, and students leave the dormitories or return home to the dormitories within the sleeping time and are classified as not attendance; the kindergarten refund report contains the use and refund conditions of various expenses in daily life, and is convenient for parents and users to check when needed;
the personnel access comprises card-free personnel access management and special personnel account management, and the card-free personnel access management comprises card-free personnel registration and card-free personnel snapshot record; when the students or teachers do not carry the campus card, the students or teachers enter or leave the campus after registering information through related personnel at the barrier-free passage group, and manual sign-in and sign-out records are recorded into attendance statistics; when an external card-free person enters the campus, the video recognition device can automatically take a snapshot record, and if the card-free person carries out check-in registration, registration information can be stored in a background database; the special personnel account management comprises grade primary account management and dormitory manager account management, and the grade primary account and the dormitory manager can log in different management interfaces respectively.
9. The campus personnel safety management system based on deep learning and face recognition features of claim 1, wherein the face information acquisition APP is used for enabling parents to register and log in to upload personal photos, the parent registration login APP is firstly downloaded to select ' simplified to people's communication APP ' to perform registration and log in, the personal photos are uploaded after entering a personal center and stored in a database for face recognition, after the uploading is completed, a video attendance check selection can be clicked to inquire the records of children entering and exiting a school, and the icons behind the corresponding records can be used for checking the photos of the children entering and exiting the school;
a teacher binds a card and uploads a personal photo, firstly downloads a ' simplified to people ' APP ' to select registration login, registers on the APP and uploads, and after completion, the teacher can upload and update student identification photos in an appointed range; in addition, a teacher can check whether students leave the school safely or not by clicking a 'safe leaving school' function, and can check identification photos of the students or teachers entering and leaving the school by clicking video attendance;
whether the school authorizes the teacher to manage the student identification photo or not needs to use the identity of a school administrator to log in background management software, authority setting in face identification photo information is selected, authority of the whole teacher or the designated teacher to manage the student identification photo is granted, and therefore the student photo can be uploaded to a background management part.
10. The deep learning and face recognition feature-based campus personnel security management system of claim 1, wherein the embedded information maintenance module comprises device information management, and firstly selects a place of a school entrance where card swiping equipment or video recognition equipment is required to be placed as a channel group, and then establishes device information management: the card swiping equipment and the video identification equipment are numbered corresponding to the campus inlet and outlet and used as a field in the system database for recording, each record indicates the position of personnel entering and exiting by using the field, and all the equipment correspond to the IP address of a local area network so as to be connected with a platform. The card swiping equipment comprises an RFID card swiping machine, a 2.4GHZ identifier and fingerprint identification equipment, wherein the RFID card swiping machine is used for identifying student cards, teacher cards and physical cards, and the 2.4GHZ identifier is mainly used for identifying 2.4GHZ active cards; the fingerprint identification device is used for fingerprint identification; the video identification device comprises a high-definition camera, an infrared sensing device and a front-end embedded identification module; the minimum requirement of the high-definition camera is 200 ten thousand pixels, and the frame rate is 25 fps; the front-end embedded identification chip comprises an ARM main control chip, a GPU chip supporting deep learning and an external monitoring camera interface, wherein the external monitoring camera interface comprises but is not limited to an MIPI interface, a DVP interface and a CSI interface.
CN201911117504.0A 2019-11-15 2019-11-15 Campus personnel safety management system based on deep learning and face recognition features Pending CN110910549A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911117504.0A CN110910549A (en) 2019-11-15 2019-11-15 Campus personnel safety management system based on deep learning and face recognition features

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911117504.0A CN110910549A (en) 2019-11-15 2019-11-15 Campus personnel safety management system based on deep learning and face recognition features

Publications (1)

Publication Number Publication Date
CN110910549A true CN110910549A (en) 2020-03-24

Family

ID=69817612

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911117504.0A Pending CN110910549A (en) 2019-11-15 2019-11-15 Campus personnel safety management system based on deep learning and face recognition features

Country Status (1)

Country Link
CN (1) CN110910549A (en)

Cited By (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111476193A (en) * 2020-04-20 2020-07-31 重庆欧啦大数据有限公司 Real-time photographing management system and method for kindergarten
CN111652058A (en) * 2020-04-27 2020-09-11 青岛百灵信息科技股份有限公司 Computer face recognition device
CN111797792A (en) * 2020-07-10 2020-10-20 重庆三峡学院 Novel identity recognition device and method based on campus management
CN111899135A (en) * 2020-07-04 2020-11-06 深圳市联想空间艺术工程有限公司 Intelligent companion chemical method and system based on face recognition
CN111931634A (en) * 2020-08-06 2020-11-13 盐城师范学院 Deep learning-based campus protection method and system
CN111932751A (en) * 2020-08-15 2020-11-13 广州云莫凡信息科技有限公司 Intelligent park Internet of things comprehensive management platform and management method
CN112102530A (en) * 2020-11-09 2020-12-18 兰和科技（深圳）有限公司 Campus Internet of things intelligent cloud lock management system
CN112288937A (en) * 2020-11-18 2021-01-29 重庆赛丰基业科技有限公司 Virtual gate and control method
CN112396714A (en) * 2020-10-30 2021-02-23 四川天翼网络服务有限公司 Non-sensing attendance system and method for school closed management
CN112418091A (en) * 2020-11-23 2021-02-26 单昂 Big data-based smart campus security data processing method
CN113012312A (en) * 2021-03-18 2021-06-22 福建省擎衣卫智能科技有限公司 Children security protection system who possesses identity recognition function
CN113361327A (en) * 2021-04-30 2021-09-07 福建榕融芯微电子科技有限公司 Campus access management system and edge computing device suitable for same
CN113554773A (en) * 2021-07-22 2021-10-26 三亚学院 Attendance monitoring device for social work practice
CN113903093A (en) * 2021-10-09 2022-01-07 福建技术师范学院 College class attendance system based on video streaming and face recognition
CN114170713A (en) * 2021-11-22 2022-03-11 浙江省邮电工程建设有限公司 Information pushing method and system based on campus management of Internet of things
CN114999017A (en) * 2022-06-06 2022-09-02 重庆酉辰戌智能科技有限公司 Campus face identification enabling system
CN115063918A (en) * 2022-04-26 2022-09-16 厦门立林科技有限公司 Face recognition method, entrance guard, intelligent lock, server and system
CN115471942A (en) * 2022-11-15 2022-12-13 内江市感官密码科技有限公司 Campus access control system and monitoring method thereof
CN116109453A (en) * 2023-01-17 2023-05-12 深圳市阳光博睿教育技术有限公司 Intelligent campus safety management early warning method and system based on artificial intelligence
CN117133069A (en) * 2023-07-25 2023-11-28 航粤智能电气股份有限公司 Temporary electric box control method based on Bluetooth lock

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101964064A (en) * 2010-07-27 2011-02-02 上海摩比源软件技术有限公司 Human face comparison method
JP2011090395A (en) * 2009-10-20 2011-05-06 Nittsu Shoji Co Ltd Security system
CN102339484A (en) * 2010-07-14 2012-02-01 新谊整合科技股份有限公司 Entrance guard and security system with function of video surveillance, and surveillance method thereof
CN104463117A (en) * 2014-12-02 2015-03-25 苏州科达科技股份有限公司 Sample collection method and system used for face recognition and based on video
CN105046795A (en) * 2015-08-28 2015-11-11 广州绿晶信息科技有限公司 Passage gate for checking attendance by photographing and working method of passage gate
CN105741209A (en) * 2016-02-04 2016-07-06 湖南亿谷科技发展股份有限公司 School management system
CN107833299A (en) * 2017-10-20 2018-03-23 上海斐讯数据通信技术有限公司 For the handheld terminal of work attendance, Time Attendance Device, attendance checking system and Work attendance method
CN109359548A (en) * 2018-09-19 2019-02-19 深圳市商汤科技有限公司 Plurality of human faces identifies monitoring method and device, electronic equipment and storage medium
CN209149373U (en) * 2018-12-29 2019-07-23 英特韦特安防科技（中山）有限公司 A kind of intelligent campus management system using theftproof card reader
CN110070469A (en) * 2019-04-18 2019-07-30 广东德融汇科技有限公司 A kind of middle and primary schools' kindergarten's Intelligent campus system based on biological identification technology

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2011090395A (en) * 2009-10-20 2011-05-06 Nittsu Shoji Co Ltd Security system
CN102339484A (en) * 2010-07-14 2012-02-01 新谊整合科技股份有限公司 Entrance guard and security system with function of video surveillance, and surveillance method thereof
CN101964064A (en) * 2010-07-27 2011-02-02 上海摩比源软件技术有限公司 Human face comparison method
CN104463117A (en) * 2014-12-02 2015-03-25 苏州科达科技股份有限公司 Sample collection method and system used for face recognition and based on video
CN105046795A (en) * 2015-08-28 2015-11-11 广州绿晶信息科技有限公司 Passage gate for checking attendance by photographing and working method of passage gate
CN105741209A (en) * 2016-02-04 2016-07-06 湖南亿谷科技发展股份有限公司 School management system
CN107833299A (en) * 2017-10-20 2018-03-23 上海斐讯数据通信技术有限公司 For the handheld terminal of work attendance, Time Attendance Device, attendance checking system and Work attendance method
CN109359548A (en) * 2018-09-19 2019-02-19 深圳市商汤科技有限公司 Plurality of human faces identifies monitoring method and device, electronic equipment and storage medium
CN209149373U (en) * 2018-12-29 2019-07-23 英特韦特安防科技（中山）有限公司 A kind of intelligent campus management system using theftproof card reader
CN110070469A (en) * 2019-04-18 2019-07-30 广东德融汇科技有限公司 A kind of middle and primary schools' kindergarten's Intelligent campus system based on biological identification technology

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
沈沛意等编著: "《DAVINCI技术剖析及实战应用开发指南》", 30 April 2012, 西安电子科技大学出版社 *
白世贞等主编: "《物联网管理学》", 30 June 2012, 中国财富出版社 *
雷鉴铭: "《ARC处理器嵌入式系统开发与编程基础》", 31 August 2019, 华中科技大学出版社 *

Cited By (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111476193A (en) * 2020-04-20 2020-07-31 重庆欧啦大数据有限公司 Real-time photographing management system and method for kindergarten
CN111652058A (en) * 2020-04-27 2020-09-11 青岛百灵信息科技股份有限公司 Computer face recognition device
CN111652058B (en) * 2020-04-27 2023-03-28 青岛百灵信息科技股份有限公司 Computer face recognition device
CN111899135A (en) * 2020-07-04 2020-11-06 深圳市联想空间艺术工程有限公司 Intelligent companion chemical method and system based on face recognition
CN111797792A (en) * 2020-07-10 2020-10-20 重庆三峡学院 Novel identity recognition device and method based on campus management
CN111931634A (en) * 2020-08-06 2020-11-13 盐城师范学院 Deep learning-based campus protection method and system
CN111932751A (en) * 2020-08-15 2020-11-13 广州云莫凡信息科技有限公司 Intelligent park Internet of things comprehensive management platform and management method
CN112396714A (en) * 2020-10-30 2021-02-23 四川天翼网络服务有限公司 Non-sensing attendance system and method for school closed management
CN112102530A (en) * 2020-11-09 2020-12-18 兰和科技（深圳）有限公司 Campus Internet of things intelligent cloud lock management system
CN112102530B (en) * 2020-11-09 2021-02-19 兰和科技（深圳）有限公司 Campus Internet of things intelligent cloud lock management system
CN112288937A (en) * 2020-11-18 2021-01-29 重庆赛丰基业科技有限公司 Virtual gate and control method
CN112418091B (en) * 2020-11-23 2021-07-13 常州易学网络科技有限公司 Big data-based smart campus security data processing method
CN112418091A (en) * 2020-11-23 2021-02-26 单昂 Big data-based smart campus security data processing method
CN113012312A (en) * 2021-03-18 2021-06-22 福建省擎衣卫智能科技有限公司 Children security protection system who possesses identity recognition function
CN113361327A (en) * 2021-04-30 2021-09-07 福建榕融芯微电子科技有限公司 Campus access management system and edge computing device suitable for same
CN113554773A (en) * 2021-07-22 2021-10-26 三亚学院 Attendance monitoring device for social work practice
CN113903093A (en) * 2021-10-09 2022-01-07 福建技术师范学院 College class attendance system based on video streaming and face recognition
CN114170713A (en) * 2021-11-22 2022-03-11 浙江省邮电工程建设有限公司 Information pushing method and system based on campus management of Internet of things
CN114170713B (en) * 2021-11-22 2023-11-21 浙江省邮电工程建设有限公司 Information pushing method and system based on campus management of Internet of things
CN115063918A (en) * 2022-04-26 2022-09-16 厦门立林科技有限公司 Face recognition method, entrance guard, intelligent lock, server and system
CN115063918B (en) * 2022-04-26 2024-01-09 厦门立林科技有限公司 Face recognition method, entrance guard, intelligent lock, server and system
CN114999017A (en) * 2022-06-06 2022-09-02 重庆酉辰戌智能科技有限公司 Campus face identification enabling system
CN115471942A (en) * 2022-11-15 2022-12-13 内江市感官密码科技有限公司 Campus access control system and monitoring method thereof
CN116109453A (en) * 2023-01-17 2023-05-12 深圳市阳光博睿教育技术有限公司 Intelligent campus safety management early warning method and system based on artificial intelligence
CN117133069A (en) * 2023-07-25 2023-11-28 航粤智能电气股份有限公司 Temporary electric box control method based on Bluetooth lock
CN117133069B (en) * 2023-07-25 2024-09-24 航粤智能电气股份有限公司 Temporary electric box control method based on Bluetooth lock

Similar Documents

Publication Publication Date Title
CN110910549A (en) Campus personnel safety management system based on deep learning and face recognition features
KR102177235B1 (en) An attendance check system using deep learning based face recognition
US10997809B2 (en) System and method for provisioning a facial recognition-based system for controlling access to a building
Samet et al. Face recognition-based mobile automatic classroom attendance management system
CN107065586B (en) Interactive intelligent home service system and method
CN105303632B (en) A kind of mobile monitor is registered system and method for work
KR20200006987A (en) Access control method, access control device, system and storage medium
CN109903414A (en) One kind is based on Internet of Things monitoring office attendance control system and method
CN110070469A (en) A kind of middle and primary schools' kindergarten's Intelligent campus system based on biological identification technology
CN116452379B (en) Intelligent campus management system based on big data
CN108153881A (en) Teaching monitoring and managing method based on Intelligent campus management
CN110188690B (en) Intelligent visual analysis system based on unmanned aerial vehicle, intelligent visual analysis system and method
CN110543811B (en) Deep learning-based non-cooperative examination personnel management method and system
CN111145414A (en) Campus safety management early warning method based on face recognition in K12 education stage
CN109522996A (en) Constant mark object automatic identification control system
CN113630721A (en) Method and device for generating recommended tour route and computer readable storage medium
CN108022191A (en) Wisdom Education Administration Information System applied to campus
KR100814510B1 (en) Entrance and exit certifying system and method thereof
KR20130109777A (en) Apparatus and method for managing attendance based on face recognition
CN109345427B (en) Classroom video frequency point arrival method combining face recognition technology and pedestrian recognition technology
CN114898443A (en) Face data acquisition method and device
CN111385530A (en) Intelligent camera combined encryption method and system
CN108538096A (en) A kind of intelligent electronic white board based on Internet of Things
CN115331346B (en) Campus access control management method and device, electronic equipment and storage medium
Park et al. A study on the design and implementation of facial recognition application system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20200324

RJ01 Rejection of invention patent application after publication