JP2005266907A - Bar code detection cutout method, device therefor and computer-readable recording medium storing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a bar code detection cutout method capable of easily and accurately detecting and cutting out a bar code from image data including a character, a pattern or the like, and to provide a device therefor and a computer-readable recording medium storing a program therefor. <P>SOLUTION: A labeling means 21 performs a labeling process to the image data 22, an effective label extraction means 23 extracts an effective label, a peak value-vicinity label extraction means 24 produces a histogram of the effective label and extracts the label having a width or a height of a peak value of the histogram or in the vicinity thereof to produce a label group, a grouping processing means 25 performs a process for regarding the label group as the same group when position relation between one label of the label group and the adjacent label is within a set range, and a bar code cutout means 26 cuts out a range to end point coordinates from start point coordinates of the group. Thereby, the bar code can be easily and accurately detected from the image data including the character, the pattern or the like to be cut out. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a barcode detection and extraction method and apparatus for detecting a barcode from image data including characters and the like, and the data, and a computer-readable recording medium storing a program.

  In general, barcodes are widely used in various fields as means for recognizing various information. When reading the barcode, there is no problem as long as the barcode portion can be directly read by the scanner, but it may be read as image information in a state where other characters and patterns are mixed. For example, a customer barcode is used for mail, and this customer barcode is read as image data together with characters such as a destination and a pattern drawn on an envelope.

  In such a case, in order to decode the barcode, it is necessary to detect and cut out the barcode portion from the image data.

  Conventionally, when the position of a barcode is detected from image data by an image processing apparatus, expansion / decompression processing for detecting the bar portion as a black lump is performed (for example, see Patent Document 1). .

However, the image expansion / reduction processing method is suitable when the size is within a certain range and the skew is small, and is not suitable for barcodes having an arbitrary number of digits, size, and arbitrary direction. Further, since each process is performed in units of one pixel, the processing time is long. For this reason, in order to shorten the processing time, dedicated image processing hardware has to be used.
JP-A-10-198755

  As described above, the conventional dilation / reduction process is suitable only for a bar code of a limited form, and there is a problem that dedicated image processing hardware is required to shorten the processing time.

  An object of the present invention is to provide a barcode detection and extraction method and apparatus capable of easily and accurately detecting and extracting a barcode from image data including characters and patterns, and a computer-readable computer storing the program Is to provide a simple recording medium.

  The barcode detection / cutout method of the present invention is a barcode detection / cutout method that detects a barcode included in image data and cuts out the data, and labels the data determined to be one connected black from the image data. The effective rectangle is extracted by comparing the bounding rectangle of this label with the reference value, and a label histogram is created from the label information of this effective label. Judge as a peak value, create a label group near the peak value by extracting a label having a height or width near this effective peak value, and the horizontal start point and vertical direction of one label in the label group near this peak value If the distance from the start point to the horizontal start point and the vertical start point of the adjacent label is within the set range, these labels are grouped together. None, the minimum x coordinate and the minimum y coordinate are obtained from the start point coordinates and the end point coordinates of all labels in the group, and are set as the start point coordinates of the barcode part, and the maximum x coordinate and the maximum y coordinate are obtained. Is the end point coordinate of the bar code part, and this start point coordinate to end point coordinate is cut out.

  In the method of the present invention, the barcode inclination angle may be calculated from the start point coordinates or end point coordinates of the foremost label in the group and the start point coordinates or end point coordinates of the end label.

  The barcode detection and clipping apparatus according to the present invention extracts a valid label from image data by comparing a labeling unit that detects one connected black image as a label and a circumscribed rectangle of the label with a reference value. An effective label extracting means for generating an effective label group, and a label histogram is created from the label information of the effective label group. If the peak value of the histogram is equal to or greater than a reference value, the effective peak value is determined. A peak value vicinity label extracting means for extracting a label having a height or width near the peak value to create a label group near the peak value, and a horizontal start point and a vertical start point of one label of the peak value vicinity label group If the distance between the horizontal start point and the vertical start point of the adjacent label is within the set range, these labels near the peak value are connected to the same group. The minimum x coordinate and the minimum y coordinate are obtained from the start point coordinates and the end point coordinates of all the labels in the group, and are set as the start point coordinates of the barcode portion, and the maximum x coordinate and the maximum A y-coordinate is obtained and used as an end point coordinate of the bar code portion, and cut-out processing means for cutting out from the start point coordinate to the end point coordinate is provided.

  Further, the apparatus of the present invention may be provided with an inclination angle calculating means for calculating the inclination angle of the barcode from the start point coordinates or end point coordinates of the foremost end label in the group and the start point coordinates or end point coordinates of the end end label.

  The computer-readable recording medium of the present invention detects, as a label, a single connected black image determined from image data, extracts the effective label by comparing the circumscribed rectangle of this label with a reference value, A label histogram is created from the label information of the effective label, and if the peak value of this histogram is greater than or equal to the reference value, it is determined as an effective peak value, and a label having a height or width near this effective peak value is extracted. A label group near the peak value is created, and the distance from the horizontal start point and the vertical start point of one label of the label group near the peak value to the horizontal start point and the vertical start point of the adjacent label is within the setting range. If so, these labels are regarded as the same group, and the minimum x coordinate and minimum y coordinate are obtained from the start point coordinates and end point coordinates of all labels in the group. This was the starting point coordinates of the bar code portion Te, which was the end point coordinate of the bar code portion for a highest x-coordinate and maximum y-coordinate is obtained by storing a program for cutting out from the start point position to the end point coordinates.

  Further, the computer-readable recording medium of the present invention is a program having a step of calculating the inclination angle of the barcode from the start point coordinates or end point coordinates of the foremost label in the group and the start point coordinates or end point coordinates of the end label. It may be stored.

  According to the present invention, the image data is labeled to create a histogram of effective labels, a label having a peak value of the histogram and a height or width in the vicinity thereof is extracted to create a label group, and this label If the positional relationship between one label in the group and the adjacent label is within the set range, it is considered as the same group, and the range from the start point coordinate to the end point coordinate of this group is cut out as a barcode part. A bar code can be detected and cut out easily from image data including the above.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a processing block diagram showing an embodiment of a barcode detection and clipping apparatus according to the present invention, and FIG. 2 is a schematic system configuration diagram for that purpose.

  In FIG. 2, reference numeral 11 denotes an image pickup unit that picks up an image of a medium 12 on which a barcode, characters, patterns, and the like are written, and outputs the image data. Reference numeral 13 denotes a general-purpose information processing apparatus, for example, a personal computer (hereinafter referred to as a personal computer), which has a printer 14 as an output device, inputs image data output from the image pickup unit 11, and generates a barcode from the image data. Detect and cut out the data.

  FIG. 1 shows processing functions in the personal computer 13. In FIG. 1, reference numeral 21 denotes a labeling unit that detects a single connected black color determined from the image data 22 from the imaging unit 11 shown in FIG.

  That is, differential labeling is performed on the entire image data 22, and it is determined as one connected black, such as a vertical bar shown in FIG. 3A, an inclined bar shown in FIG. A horizontal bar shown in FIG. 3C, characters shown in FIG. 4D, and the like are detected as labels. A group of these detected labels is referred to as a label group A.

  For each of these labels, the start point coordinates (xs, ys) and end point coordinates (xe, ye), height: h, and width: w of the circumscribed rectangle 19 are obtained as label information. Note that the start point coordinates (xs, ys) and the end point coordinates (xe, ye) are obtained on the basis of the origin coordinates (x0, y0) in the upper left of the image memory (not shown) of the personal computer 13 as a reference. Further, the height: h and the width: w are calculated based on the start point coordinates (xs, ys) and the end point coordinates (xe, ye).

  An effective label extracting means 23 compares the circumscribed rectangle 19 of the label with a reference value, extracts an effective label, and creates an effective label group B.

  That is, the bar height of the bar code is normally 6 mm or more, and even if it is inclined as shown in FIG. 3B, it should protrude from a 4 mm square. Therefore, as shown in FIG. 3, a 4 mm square 18 is compared as a reference value, only labels having a size of 4 mm square or less are extracted as effective labels, and a serial number is assigned to each label to provide an effective label. Group B is created. Note that labels less than 4 mm square are determined to be invalid labels and discarded.

  Reference numeral 24 denotes a peak value vicinity label extracting means. First, a label histogram is created from the label information of the effective label group B, and each peak value is calculated. If the peak value of the histogram is equal to or greater than the reference value, it is determined that the peak value is effective, and a label having a height or width near the effective peak value is extracted to create a peak value vicinity label group C.

  That is, within the same bar code, the heights of the bars are substantially the same, and many are arranged in the same direction. Even if the case of tilting is taken into consideration, if each histogram is taken for height and width, the label of the barcode bar will be included in the label label having the height or width near each peak value of the histogram. Therefore, a label having a height or width near the peak value is extracted.

  For example, when the height of the barcode is 6 mm, if the barcode is upright, the length data having a height of 6 mm has the largest peak value, and data near the peak value is also extracted. In this case, if the peak value is less than 4 mm, it is ignored as an invalid peak value.

  Here, if the effective peak value is only the height, it is determined as an erect barcode, and only the label having a height near the height peak value is extracted. If only the width peak value is valid, it is determined that the bar code is an inverted barcode, and only labels having a width near the width peak value are extracted. Furthermore, if both peak values are valid, it is determined that the barcode is slanted, and if there is a height near the height peak value, both a label and a label with a width near the width peak value are extracted. . And the peak value vicinity label group C is created with these extracted labels.

  Reference numeral 25 denotes grouping processing means. If both the horizontal start point and the vertical start point of one label in the label group near the peak value to the horizontal start point and vertical start point of the adjacent label are within the set range, these peak values Neighboring labels are considered as the same group. That is, paying attention to the fact that the barcode bars are adjacent to each other, the adjacent label grouping process is performed on the peak value vicinity label group C.

  For example, as shown in FIG. 4, when the barcode 27 is inclined, the horizontal start point and the vertical start point of one label 27-1 of the label group C near the peak value and the horizontal direction of the adjacent label 27-2. It is determined whether the difference between the direction start point and the vertical direction start point is within a preset range MAXDES. As a result, if each is within the set range MAXDES, these labels 27-1 and 27-2 are regarded as the same group and are given the same group number. Since the value of the setting range MAXDES varies depending on the resolution of the captured image and the maximum width of the barcode to be detected, it is treated as a variable variable.

  This grouping process is performed for all the labels 27-1, 27-2,..., 27-12 of the label group C near the peak value, and finally the group in which the number of labels in the group is equal to or greater than the set value MINBARCOUNT. Is determined as a barcode candidate. The set value MINBARCOUNT is treated as a variable variable because the optimum value differs depending on the type of barcode to be detected and the minimum number of digits.

  26 is a cut-out processing means for cutting out from the start point coordinates to the end point coordinates of the range including all the labels in the group determined as the barcode candidate as a barcode part.

  That is, the label information of the label group that is set as the barcode candidate is sorted. If it is determined to be upright or inclined, it is sorted in ascending order in the horizontal direction, and if it is determined to be inverted, it is sorted in descending order in the vertical direction, and the position information of the barcode candidates from the information of the first label and the last label Create

  For example, as shown in FIG. 5, if the minimum x coordinate and the minimum y coordinate are obtained from the start point coordinates of all labels in the group of barcode candidates, they become the start point coordinates (sx1, sy2) of the barcode candidate. If the maximum x coordinate and the maximum y coordinate are obtained from the end point coordinates, they become the end point coordinates (ex2, ey1) of the barcode candidate. The range from the start point coordinates to the end point coordinates is the inspection cutout result.

  Reference numeral 28 denotes an inclination angle calculation means for calculating the inclination angle of the barcode from the start point coordinates or end point coordinates of the foremost end label 27-1 in the group and the start point coordinates or end point coordinates of the end end label 27-12.

That is, as shown in FIG. 5, the barcode inclination angle θ from the start point coordinates (sx1, sy1) of the foremost label 27-1 and the start point coordinates (sx2, sy2) of the end label 27-12 in the group. Is calculated by the following equation.

  Of course, it may be calculated from the end point coordinates (ex1, ey1) and the end point label 27-12 end point coordinates (ex2, ey2) of the foremost end label 27-1.

  By obtaining the inclination angle θ of the barcode 27 in this way, the barcode can be established by subjecting the image data of the cut-out portion to an angle θ rotation process. As a result, the processing time in the subsequent recognition process can be shortened.

  Next, the flow of processing will be described using the flowchart of FIG. When the barcode detection / cutout process is started for the image data 21 including the barcode, first, the labeling process is performed on the entire image (601). By this processing, what is determined as one connected black in the image is detected as one label, and a label group A is created. As the label information, as shown in FIG. 3, information of the start point coordinate, end point coordinate, height h, and width w of the circumscribed rectangle 19 is obtained for each label.

  Next, in comparison with a reference value (in this example, a range of 4 mm square) 18, those exceeding the reference value are extracted as effective labels to create an effective label group B (602). The number of labels of this effective label group is compared with the set value MINBARCOUNT (603). As a result, if the number of labels does not reach the set value MINBARCOUNT, it means that there is no barcode, and the process ends.

  On the other hand, if the number of labels is equal to or greater than the set value MINBARCOUNT, a histogram for each of the label height and width is created from the label information of the effective label group B, and the peak value is calculated (604). Among these peak values, those less than the reference value (4 mm in this example) are discarded as invalid, labels having valid peak values and values near the peak values are extracted, and the peak value vicinity label group C is selected. Create (605).

  Next, the labels having the relationship shown in FIG. 4 are grouped for the label group C near the peak value (606). Then, the group where the number of labels in the group is equal to or greater than the set value MINBARCOUNT is set as a barcode candidate, and the position of the barcode candidate is searched by the processing of FIG. 5 (607) and cut out.

  If there are a plurality of barcodes to be detected and cut out, they are compared with the preset maximum number of discoveries MAXCODECOUNT (608). Then, it is deleted from the effective label group B (609). The number of labels after deletion is compared with the set value MINBARCOUNT (603), and if it is less than the set value MINBARCOUNT, the process is terminated because there is no more barcode. Even after the deletion, if the number of labels is the set value MINBARCOUNT, the processes after the histogram creation (604 to 609) are repeated again.

  As a result, all the various barcodes (for example, a plurality of barcodes having different heights) existing in the image data can be detected.

  In this manner, a plurality of different types of barcodes can be detected and cut out accurately and in a short time by means that are more versatile than conventional methods and are inexpensive.

It is a functional block diagram explaining one Embodiment of the barcode detection clipping apparatus by this invention. It is a system conceptual diagram of one embodiment same as the above. (A) (b) (c) (d) is a figure explaining the determination method of the effective label by one Embodiment of the barcode detection clipping method by this invention, respectively. It is a figure explaining the label grouping method by one Embodiment of the barcode detection clipping method by this invention. It is a figure explaining the cutout range by one Embodiment of the barcode detection cutout method by this invention. It is a flowchart explaining the process sequence in one Embodiment of the barcode detection clipping method by this invention.

Explanation of symbols

21 image data,
22 Labeling means 23 Effective label extracting means 24 Peak value vicinity label extracting means 25 Grouping processing means 26 Bar code cutting means 27 Bar code 28 Inclination angle calculating means

Claims (6)

A barcode detection and clipping method for detecting a barcode included in image data and cutting out the data,
From the image data, one that is judged to be connected black is detected as a label, and the bounding rectangle of this label is compared with a reference value to extract an effective label,
A label histogram is created from the label information of this effective label. If the peak value of this histogram is equal to or greater than the reference value, it is determined as an effective peak value, and a label having a height or width near this effective peak value is extracted. To create a label group near the peak value,
If the distance from the horizontal start point and vertical start point of one label in the label group near this peak value to the horizontal start point and vertical start point of the adjacent label is within the set range, these labels are regarded as the same group. ,
From the start point coordinates and end point coordinates of all the labels in the group, the minimum x coordinate and the minimum y coordinate are obtained and used as the start point coordinates of the barcode part, and the maximum x coordinate and the maximum y coordinate are obtained to obtain this bar code. A bar code detection cutout method characterized in that the end point coordinates of a code part are used and the start point coordinates to the end point coordinates are cut out.   2. The barcode detection cutout method according to claim 1, wherein the inclination angle of the barcode is calculated from the start point coordinates or end point coordinates of the foremost end label in the group and the start point coordinates or end point coordinates of the end end label. A barcode detection and extraction device that detects a barcode included in image data and extracts the data,
Labeling means for detecting, as a label, one that is determined to be one connected black from the image data;
An effective label extraction means for extracting an effective label by comparing a circumscribed rectangle of the label with a reference value and creating an effective label group;
A label histogram is created from the label information of the effective label group. If the peak value of the histogram is equal to or greater than a reference value, it is determined as an effective peak value, and a label having a height or width near the effective peak value is determined. A peak value vicinity label extracting means for extracting and creating a peak value vicinity label group,
If the distance from the horizontal start point and the vertical start point of one label in the label group near the peak value to the horizontal start point and the vertical start point of the adjacent label is within the setting range, the labels near the peak value are the same. Grouping processing means regarded as a group,
From the start point coordinates and end point coordinates of all the labels in the group, the minimum x coordinate and the minimum y coordinate are obtained and used as the start point coordinates of the barcode part, and the maximum x coordinate and the maximum y coordinate are obtained to obtain this bar code. Cutout processing means for cutting out from the start point coordinates to the end point coordinates as the end point coordinates of the code part,
A bar code detection and clipping device comprising:   The inclination angle calculation means for calculating the inclination angle of the barcode from the start point coordinates or end point coordinates of the foremost end label in the group and the start point coordinates or end point coordinates of the end end label is provided. Barcode detection clipping device. From the image data, one that is judged to be connected black is detected as a label, and the bounding rectangle of this label is compared with a reference value to extract an effective label,
A label histogram is created from the label information of this effective label. If the peak value of this histogram is equal to or greater than the reference value, it is determined as an effective peak value, and a label having a height or width near this effective peak value is extracted. To create a label group near the peak value,
If the distance from the horizontal start point and vertical start point of one label in the label group near this peak value to the horizontal start point and vertical start point of the adjacent label is within the set range, these labels are regarded as the same group. ,
From the start point coordinates and end point coordinates of all the labels in the group, the minimum x coordinate and the minimum y coordinate are obtained and used as the start point coordinates of the barcode part, and the maximum x coordinate and the maximum y coordinate are obtained to obtain this bar code. A computer-readable recording medium that stores a program that takes the end point coordinates of the code portion and cuts from the start point coordinates to the end point coordinates.   6. The computer-readable record according to claim 5, wherein a program having a step of calculating an inclination angle of a bar code from a start point coordinate or end point coordinate of a front end label in a group and a start point coordinate or end point coordinate of a rear end label is stored. Medium.

Images