JP4653215B2 - Code recognition method and apparatus, electronic apparatus, program, and storage medium - Google Patents

Description

  The present invention relates to a code recognition method and apparatus, an electronic apparatus, a program, and a storage medium, and in particular, a code recognition method and apparatus for recognizing a one-dimensional code or a two-dimensional code included in image data, an electronic apparatus, a program, and a computer readable medium. The present invention relates to a storage medium.

  FIG. 1 is a diagram illustrating an example of a one-dimensional barcode and a two-dimensional code. In recent years, two-dimensional codes that can embed more information than conventional one-dimensional barcodes have become popular. In the two-dimensional code, the code cannot be recognized unless the two-dimensional array data of the entire code is read. For this reason, an image reading device such as a digital camera or an image scanner is used for reading the two-dimensional code.

  With the widespread use of two-dimensional codes, a method for automatically recognizing both one-dimensional codes and two-dimensional codes using a single image reading apparatus has been proposed. Patent Document 1 proposes a method for recognizing a code by identifying a block including a one-dimensional or two-dimensional code that divides image data of a read code into rectangular blocks and scans as needed and satisfies a predetermined condition. Yes. However, these code recognition methods recognize a code printed in black, which has been widely used in the past, and the image data used for recognition may be a single, that is, a gray-scale color plane. It is a premise.

  With the spread of two-dimensional codes, the use of codes printed in colors other than black is increasing from the viewpoint of design and the like. However, since the conventional code recognition method recognizes code using a single (grayscale) color plane image in accordance with the assumption that the coat to be recognized is black, it was created in a color other than black. The code is difficult to recognize.

  The problem in the case of recognizing a code printed in a color other than black using a gray scale color plane will be described more specifically. The gradation values of the gray scale (Y), red (R), green (G), and blue (B) color planes generally have the following relationship.

Y = 0.299 × R + 0.587 × G + 0.114 × B
Normally, the code is printed in black (R = G = B = 0) and the background is printed in white (R = G = B = 255), so the tone difference between the code in Y and the background is 255. Here, when the code is printed in red (R = 255, G = B = 0), the contribution of the code to Y is 29% (Y = 76), and the tone difference between the code and the background in Y Decreases to 179. As the tone difference decreases in this way, it is difficult to accurately recognize the code.
JP 2005-173646 A

  As described above, since the conventional code recognition method is considered on the assumption that the code to be read is black, there is a problem that it is difficult to accurately recognize codes of colors other than black.

  However, since the conventional code recognition method has started from the recognition of a one-dimensional bar code printed in black and has evolved into the recognition of a two-dimensional code, it is not assumed that codes of colors other than black will be read. Conventionally, a recognition technique itself for recognizing codes other than black has not been proposed.

  Accordingly, an object of the present invention is to provide a code recognition method and apparatus, an electronic apparatus, a program, and a storage medium that can accurately recognize a code regardless of the color of the code.

  The problems described above include a step of dividing the color image data of the read code into color components of a plurality of color planes, a selection step of selecting one color plane, and a code based on the color component data of the selected color plane. And a recognition step for recognizing the code.

  The above problem is that the color image data of the read code is divided into color components of a plurality of color planes, a selection unit that selects one color plane, and a code based on the color component data of the selected color plane. It can also be achieved by a code recognition method characterized by comprising recognition means for recognizing.

  The above object can also be achieved by an electronic device comprising an image reading device for reading a code and the code recognition device as described above.

  The above object can also be achieved by a program including a procedure for causing a computer to operate as the code recognition device as described above.

  The above problem can also be achieved by a program including a procedure for causing a computer to operate as the electronic device as described above.

  The above object can also be achieved by a computer-readable storage medium characterized by storing the above program.

  According to the present invention, it is possible to realize a code recognition method and apparatus, an electronic apparatus, a program, and a storage medium capable of accurately recognizing a code regardless of the color of the code.

It is a figure which shows an example of a one-dimensional barcode and a two-dimensional code. It is a block diagram which shows the principal part of one Example of the electronic device which becomes this invention. It is a flowchart explaining operation | movement of one Example of this invention. It is a figure explaining extraction of a block. It is a flowchart explaining calculation of the feature-value of a block in detail about a two-dimensional code. It is a figure explaining the number of horizontal direction edges and the number of vertical direction edges of a one-dimensional code. It is a figure explaining the number of horizontal direction edges and the number of vertical direction edges of a two-dimensional code. It is a flowchart explaining calculation of the feature-value of a block in detail about a one-dimensional code. It is a figure explaining the correlation between horizontal lines and the correlation between vertical lines of a one-dimensional code. It is a figure explaining the correlation between horizontal lines and the correlation between vertical lines of a two-dimensional code. It is a figure explaining code recognition in case a code and other information are piled up. It is a block diagram explaining the other Example of the code recognition apparatus which becomes this invention.

Explanation of symbols

1 Electronic device 9 Image reading device 10 CPU
11 ROM
12 RAM
13 Communication interface 14 Storage device 15 Input / output device 16 Media drive 18 Information provider 19 Network 20 Recording medium 31 Two-dimensional code 41 One-dimensional code 51 Code recognition device

  In the present invention, the image data of the read code is divided into color components of a plurality of color planes, and the code is recognized based on the color component data of the selected color plane. The color plane to be selected may be determined in advance, or a color plane having a large feature amount in the code or a small gradation value may be automatically selected. Particularly in the latter case, even if the code has different image characteristics depending on the color plane, the code can be reliably and accurately recognized. Further, based on the feature amount, it may be determined whether the code is a one-dimensional code or a two-dimensional code.

  When the color used for the code to be read is known in advance, the number of color planes to be divided may be two or more.

  On the other hand, when the color used for the code to be read is not known in advance, the number of color planes to be divided is 3 or more, preferably at least three primary color planes, so that the code can be recognized reliably. Can be done. In this case, it is possible to automatically select a color plane suitable for the color of the read code and perform recognition processing, so that any color code can be recognized reliably and accurately. It becomes.

  Embodiments of a code recognition method and apparatus, electronic apparatus, program, and storage medium according to the present invention will be described below with reference to FIG.

  FIG. 2 is a block diagram showing the main part of an embodiment of the electronic apparatus according to the present invention. The electronic device 1 shown in FIG. 1 includes an image reading device 9, a CPU 10, a ROM 11, a RAM 12, a communication interface 13, a storage device 14, an input / output device 15, and a medium drive 16 connected by a bus 17.

  This embodiment of the electronic apparatus employs an embodiment of a code recognition method and apparatus, program, and storage medium according to the present invention.

  The image reading device 9 includes a digital camera and an image scanner. The CPU 10 governs overall control of the electronic device 1. The ROM 11 stores programs and data executed by the CPU 10. The RAM 12 provides a work area and a data storage area when the CPU 10 executes a program. The communication interface 13 can be connected to the information provider 18 via the network 19. The information provider 18 may be another electronic device or its storage device. The storage device 14 includes a semiconductor storage device, a magnetic storage device (for example, a magnetic disk device), an optical storage device (for example, an optical disk device), a magneto-optical storage device (for example, a magneto-optical disk device), and the like. The input / output device 15 includes an input unit configured with a keyboard or the like, and an output unit configured with a liquid crystal display panel or the like. The medium drive 16 writes information to and / or reads information from a portable recording medium 20 such as a disk. The ROM 11 constitutes a readable storage device of the electronic device, and the RAM 12, the storage device 14, and the medium drive 16 loaded with the recording medium 20 constitute a readable / writable storage unit of the electronic device 1.

  The electronic device 1 may be a desktop personal computer, a laptop (portable) personal computer, an information processing device such as a PDA, or a communication device such as a mobile phone. When the electronic device 1 is a mobile phone, the communication interface 13 is a transmission / reception unit that communicates with the base station and the other party mobile phone via the mobile communication network.

  This embodiment of the code recognition device is constituted by at least the CPU 10 in the electronic device 1. This embodiment of the code recognition method and program is executed by the CPU 10. The present embodiment of the storage medium is configured by a computer-readable storage medium storing the present embodiment of the program. The computer-readable storage medium can be configured by a readable storage unit, and is configured by, for example, the recording medium 20. This embodiment of the program may be stored in a readable / writable storage unit in the electronic apparatus 1 from the information provider 18 via the network 19 and the communication interface 13.

  When recognizing the one-dimensional code or the two-dimensional code, the code is read by the image reading device 9 by a known method, and the image data is stored in a readable / writable storage unit. In this case, the code to be read may be printed on a medium such as paper or may be displayed on a display screen. Since the code is read by the image reading device 9 such as a digital camera, it is difficult to read the image of only the code. Usually, the image data of the read code is the image data related to the code image and the surroundings of the code. It consists of image data relating to an image of a certain background.

  FIG. 3 is a flowchart for explaining the operation of this embodiment. The processing shown in the figure is executed by the CPU 10 when the code is read and the image data is stored in a readable / writable storage unit. Here, for convenience of explanation, it is assumed that the two-dimensional color code is read and the color image data is stored in a readable / writable storage unit. In FIG. 3, step S1 divides color image data stored in a readable / writable storage unit into a plurality of arbitrary color components, and creates a color plane for each color component. In this embodiment, for convenience of explanation, color image data is divided into three primary color components of red (R), green (G), and blue (B) to create an R plane, a G plane, and a B plane. Shall.

  In step S2, in each color plane created in step S1, a rectangular area (hereinafter referred to as a block) of interest is sequentially extracted by scanning the image sequentially from an arbitrary position of the image. The block can be selected by scanning from the upper left corner of the image in the left and right main scanning directions and proceeding in the upper and lower sub scanning directions, or by scanning from the center of the image toward the periphery. There is no particular limitation on the scanning start position and scanning direction. In order to detect an image of the read code in a short time, it is desirable to start scanning from a position in the image where there is a high possibility that the code exists and extract a block.

  FIG. 4 is a diagram for explaining block extraction. This figure shows a case where the read two-dimensional color code 31 is red for convenience of explanation. The center part of FIG. 4 shows a state where blocks 311R, 311G, and 311B of interest are extracted from the center of the image as an example in each of the R, G, and B planes 31R, 31G, and 31B. Since the read two-dimensional color code 31 is red, each pixel is represented by 256 gradation levels (for example, “255” indicates the highest luminance and “0” indicates the lowest luminance). The minimum value of the pixel data in the block 311R in the R plane 31R is “255”, the minimum value of the pixel data in the block 311G in the G plane 31G is “0”, and the minimum value of the pixel data in the block 311B in the B plane 31B Is “10”.

  In step S3, the feature amount of the code portion of the two-dimensional color code included in the extracted block is calculated. Discrimination between the code portion and the background portion in the extracted block will be described later. Here, the code portion is obtained by using the minimum value of the pixel data in the block 311R in the R plane 31R, the minimum value of the pixel data in the block 311G in the G plane 31G, and the minimum value of the pixel data in the block 311B in the B plane 31B. The feature amount is calculated. When the minimum value of the pixel data is small, it means that the difference in contrast with the background in which the pixel indicated by the pixel data is low, that is, the pixel is dark and the pixel luminance is usually high can be increased.

  Both the one-dimensional code and the two-dimensional code are printed using two colors having a large contrast difference such as a combination of white and black. Since codes are usually printed using white as the background color, the code color is darker, that is, the smaller the pixel value, the greater the difference in contrast from the background color, and the code is a unified code. It is possible to accurately determine whether the code is a two-dimensional code and recognize the code.

  In step S4, the feature amount calculated in step S3 is compared between the R, G, and B planes 31R, 31G, and 31B, and a color plane in which the feature amount is strongly expressed is selected for subsequent processing. In this example, the G plane 31G having the smallest feature value is selected from among the pixel data in the blocks 311R, 311G, and 311B. Since the two-dimensional color code is red, in this case, the feature quantity does not appear on the R plane 31R, while the feature quantity appears on the G and B planes 31G and 31B. Here, in order to use the image data with the highest contrast, the G plane 31G having the larger feature amount is selected.

  In step S5, recognition processing such as discrimination between a one-dimensional code and a two-dimensional code and recognition of code contents is performed using the image data of the two-dimensional code 31 of the selected G plane 31G. Since the recognition process is performed using the image data of the G plane 31G having the largest feature amount, the recognition process can be performed with higher accuracy.

  Note that the selection of the color plane in step S4 is not limited to the method of selecting the color plane having the largest feature amount. For example, the color plane having the highest preset priority order may be selected from among the R, G, B planes 31R, 31G, 31B having a minimum value of pixel data of the extracted block equal to or less than a predetermined value. When the color plane is selected in this way, for example, even if a blue character or the like is written on the red two-dimensional code, the B plane 31B is selected and used for the recognition processing in step S5, so that the blue character or the like is selected. It is possible to perform highly accurate recognition processing without being affected by the above. However, in this case, the B, G, and R planes 31B, 31G, and 31R are preset in descending order of priority, and the minimum value of the pixel data of the extracted block is, for example, among “15” or less color planes. It is assumed that the color plane with the highest priority is selected. In this way, the color plane selection criterion in step S4 may be selected in accordance with the state of the color code and the processing performed in the recognition processing in step S5.

  In the recognition processing in step S5, if it is known in advance that the color code to be read is a two-dimensional color code, the code content is recognized by a known method using the image data of the color plane selected in step S4. be able to. If it is not known beforehand whether the color code to be read is a one-dimensional color code or a two-dimensional color code, the color code read using the image data of the color plane selected in step S4. It is also possible to recognize the code contents after determining whether is a one-dimensional color code or a two-dimensional color code.

  FIG. 5 is a flowchart for explaining the calculation of the block feature amount in step S3 shown in FIG. 3 in more detail for the two-dimensional code. In FIG. 5, the same steps as those in FIG. 3 are denoted by the same reference numerals, and the description thereof is omitted.

  In FIG. 5, step S31 calculates the feature amount of the block extracted from the read code as described above. In step S32, for example, for the G plane 31G having the smallest minimum pixel value, it is determined whether or not the standard deviation SD of the pixel value of the extracted block 311G is larger than the threshold value SDc. Both the one-dimensional code and the two-dimensional code are printed using two colors having a large contrast difference such as a combination of white and black. In the code portion (region in the code), the area ratio occupied by the two colors is close to 1: 1. For this reason, the code portion has a feature that the variation (or standard deviation) of the pixel value is larger than the variation (or standard deviation) of the pixel value in the background portion. Thus, since the standard deviation SD of the pixel value is different between the code portion and the background portion, it can be determined that the code portion is the code portion if the standard deviation SD is larger than the threshold value SDc. In addition, as the contrast difference is larger, it is possible to accurately determine whether the code is a unitary code or a two-dimensional code and to recognize the code. Therefore, by selecting a color plane of a block in which the standard deviation SD of the pixel values in the block is larger than the threshold value SDc and has the maximum standard deviation SD among such standard deviations SD, the code is a unified code. Or two-dimensional code and code recognition can be performed with higher accuracy.

  If the decision result in the step S32 is YES, a step S33 decides whether, for example, the horizontal edge number He of the extracted block 311G is larger than the threshold value Emin and the vertical edge number Ve is larger than the threshold value Emin. .

  The one-dimensional code is a code that represents data by a combination of parallel straight lines (bars) having different thicknesses and intervals. When pixels on a straight line perpendicular to the bar are taken out, there are as many change points (edges) of the bar pixel value as there are bars corresponding to the number of bars included on the extracted straight line, regardless of where the code portion is cut out. When pixels on a straight line parallel to the bar are taken out, there is a feature that no edge exists. On the other hand, the two-dimensional code is a code that represents data by a combination of light and dark dots (cells) arranged two-dimensionally. The cells are configured to be arranged as uniformly as possible in the code portion. When pixels on a straight line parallel to and perpendicular to the cell array are taken out, there is a feature that the change point of the pixel value exists on both the horizontal and vertical lines regardless of where the code portion is cut out. . Therefore, by selecting a color plane having the largest number of edges in the horizontal and vertical lines and having the largest block, it is possible to determine whether the code is a one-dimensional code or a two-dimensional code. Code recognition can be performed more accurately.

  FIG. 6 is a diagram illustrating the number of horizontal edges He and the number of vertical edges Ve of a one-dimensional code, and FIG. 7 is a diagram illustrating the number of horizontal edges He and the number of vertical edges Ve of a two-dimensional code. is there. The horizontal edge number He is an edge of a pattern that traverses when the code portion in the extracted block is scanned in the horizontal direction (that is, the change point of the pixel value), and the vertical edge number Ve is the extracted block. This is an edge of a pattern that traverses when the inner code portion is scanned in the vertical direction (that is, a change point of pixel value). 6 and 7, the upper part shows the entire code, and the lower part shows the extracted block, the number of horizontal edges He, and the number of vertical edges Ve. As can be seen from FIG. 6, in the case of a one-dimensional code, the horizontal edge number He is the same as “10” regardless of the vertical position indicated by the wavy line, and the vertical edge number Ve is the same. All the scans at any position in the horizontal direction indicated by the broken line are “0”. On the other hand, as can be seen from FIG. 7, in the case of the two-dimensional code, the horizontal edge number He differs from “4”, “6”, “7” depending on the position indicated by the wavy line, and the vertical edge number Ve is It differs from “7”, “2”, and “4” depending on the position indicated by the wavy line. In the case of a two-dimensional code, since the code pattern is configured such that the number of horizontal edges He and the number of vertical edges Ve in a block of a predetermined size are both greater than a predetermined threshold value Emin, the one-dimensional code is converted in step S33. A two-dimensional code can be distinguished.

  If the decision result in the step S33 is YES, a step S34 decides whether or not the black pixel rate BPr of the extracted block is larger than the lower limit threshold BPmin and smaller than the upper limit threshold BPmax. Here, it is assumed that the reference value BPR of the black pixel ratio BPr is included in the range between the lower limit threshold BPmin and the upper limit threshold BPmax. The black pixel ratio BPr is a ratio of pixels having a small pixel value (for example, the minimum luminance value) to a pixel having a large pixel value (for example, the maximum luminance value). Both the one-dimensional code and the two-dimensional code are printed using two colors having a large contrast difference such as a combination of white and black. Assuming that the background portion is a white pixel and the code portion is a black pixel, particularly in the two-dimensional code portion, the area ratio between the white pixel and the black pixel includes a predetermined range (for example, 50%) (for example, , 50% ± 10%), the code pattern is configured in advance. Thus, since the black pixel rate is different between the code portion and the background portion, if the black pixel rate BPr is larger than the lower threshold BPmin (> BPR) and smaller than the upper threshold BPmax (<BPR), the code portion, that is, It can be determined that it is a code portion of the two-dimensional code (since the determination result in step S33 is YES). Therefore, the code is unified by selecting a color plane having a block whose black pixel ratio (or white pixel ratio) in the block is within a predetermined range including the reference value and having a block closest to the reference value. Whether it is a code or a two-dimensional code and code recognition can be performed with higher accuracy.

  If the determination result of any of steps S32, S33, and S34 is NO, step S35 determines that there is no two-dimensional code in the extracted block, and the process returns to step S2 to extract the next block. On the other hand, if the decision result in the step S34 is YES, a step S36 judges that there is a two-dimensional code in the extracted block, and the process advances to a step S4. In this case, since it is determined that the color code read in step S36 is a two-dimensional color code, in step S5, the code content of the two-dimensional code using the image data of the color plane selected in step S4. May be recognized by a known method.

  If the process proceeds to step S35 even after a predetermined number of blocks have been extracted, for example, even if an error message is output to the input / output unit 15, it is determined whether there is a one-dimensional code described later with reference to FIG. May be performed.

  FIG. 8 is a flowchart for explaining the calculation of the block feature amount in step S3 shown in FIG. 3 in more detail for the one-dimensional code. In FIG. 8, the same steps as those in FIG. 3 are denoted by the same reference numerals, and the description thereof is omitted.

  In FIG. 8, step S301 calculates the feature amount of the block extracted from the read code as described above. In step S302, for example, for the G plane 31G having the smallest pixel value, it is determined whether or not the standard deviation SD of the pixel value of the extracted block 311G is larger than the threshold value SDc. Since the standard deviation SD of the pixel value is different between the code portion and the background portion, if the standard deviation SD is larger than the threshold value SDc, it can be determined that it is the code portion. If the decision result in the step S302 is YES, a step S303 decides whether, for example, the horizontal edge number He of the extracted block 311G is larger than the threshold value Emin or whether the vertical edge number Ve is larger than the threshold value Emin. To do. As described with reference to FIGS. 6 and 7, since the number of vertical edges Ve is “0” in the case of a one-dimensional code, it is determined whether or not there is a code portion of the one-dimensional code in the block extracted in step S303. be able to.

  If the decision result in the step S303 is YES, a step S304 decides whether or not the horizontal line correlation Hs of the extracted block is larger than the threshold value Smin or whether the vertical line correlation Vs is larger than the threshold value Smin.

  The one-dimensional code shows an extremely strong correlation when a cross-correlation is taken between a straight line perpendicular to the bar and the parallel line in the vicinity of the code part. When cross-correlation is taken, there is a characteristic that there is almost no correlation. On the other hand, the two-dimensional code has a feature that there is no particularly clear correlation when a cross-correlation is taken between a straight line parallel and perpendicular to the cell arrangement and a parallel straight line in the vicinity thereof. Also, in the case of a background portion where characters, photographs, etc. are printed, the correlation is not as strong as that of a one-dimensional code. Accordingly, by selecting a color plane having a block in which the correlation between specific horizontal lines in a block and the correlation between specific vertical lines are within a predetermined range, and the code is the largest, the code is a unitary code. Or two-dimensional code and code recognition can be performed with higher accuracy.

  FIG. 9 is a diagram illustrating the horizontal line correlation Hs and the vertical line correlation Vs of the one-dimensional code, and FIG. 10 is a diagram illustrating the horizontal line correlation Hs and the vertical line correlation Vs of the two-dimensional code. is there. The inter-horizontal line correlation Hs is a cross-correlation pattern when the code portion in the extracted block is scanned in the horizontal direction, and the vertical inter-line correlation Vs is scanned in the code portion in the extracted block in the vertical direction. It is the correlation of the pattern that crosses the case. 9 and 10, the upper part shows the entire code, and the lower part shows the extracted block, horizontal line correlation Hs, and vertical line correlation Vs. As can be seen from FIG. 9, in the case of a one-dimensional code, the horizontal line correlation Hs is large regardless of the vertical position indicated by the one-dot chain line, and the horizontal line correlation Vs is indicated by the one-dot chain line. There is no scanning in the direction position. On the other hand, as can be seen from FIG. 10, in the case of a two-dimensional code, the horizontal line correlation Hs is small at any position indicated by the one-dot chain line, and the vertical line correlation Vs is indicated by the one-dot chain line. It is small even if it scans with. In the case of a one-dimensional code, the code pattern is configured such that the horizontal line correlation Hs in the block is large. Therefore, if the horizontal line correlation Hs or the vertical line correlation Vs is greater than the threshold value Smin, It can be determined that it is a code portion (since the determination result in step S303 is YES).

  If the determination result in any of steps S302, S303, and S304 is NO, step S305 determines that there is no one-dimensional code in the extracted block, and the process returns to step S2 to extract the next block. On the other hand, if the decision result in the step S304 is YES, a step S306 judges that there is a one-dimensional code in the extracted block, and the process advances to a step S4. In this case, since it is determined that the color code read in step S306 is a one-dimensional color code, in step S5, the code content of the one-dimensional code using the image data of the color plane selected in step S4. May be recognized by a known method.

  If the process proceeds to step S305 after extracting a predetermined number of blocks, for example, even if an error message is output to the input / output unit 15, it is determined whether or not there is the two-dimensional code described with reference to FIG. May be performed.

  When it is known that the code to be read is a two-dimensional code, only the processing of FIG. 5 needs to be performed. When it is known that the code to be read is a one-dimensional code, FIG. It is sufficient to perform only the process.

  If it is not known whether the code to be read is a one-dimensional code or a two-dimensional code, the process of FIG. 5 is performed and step S35 is performed a predetermined number of times. Even if the process is switched to the process of FIG. 8, when the process of FIG. 8 is performed and step S305 is performed a predetermined number of times, the process of step S3 may be switched to the process of step S3 shown in FIG. Furthermore, it is possible to perform the processes of step S3 shown in FIG. 5 and step S3 shown in FIG. 8 in parallel. Since the code part of the one-dimensional code and the two-dimensional code has a plurality of feature quantities different from the background part, whether the code is a unified code by selecting a color plane by combining these feature quantities. Whether it is a two-dimensional code and code recognition can be performed with higher accuracy.

  FIG. 11 is a diagram for explaining code recognition when a code and other information are superimposed. For the sake of convenience of explanation, the figure shows a case where the read one-dimensional color code 41 is black and a red character “FUJITSU” is written on the code portion. The center part of FIG. 11 shows a state in which the blocks 411R, 411G, and 411B of interest are extracted from the right part from the center of the image in each R, G, B plane 41R, 41G, 41B as an example. Since characters are written on the code portion of the read one-dimensional color code 41, the horizontal line correlation Hs of the pixel data in the block 411R in the R plane 41R is “0.90” as an example, and the G plane. The inter-horizontal line correlation Hs of the pixel data in the block 411G in 41G is “0.20”, and the inter-horizontal line correlation Hs of the pixel data in the block 411B in the B plane 41B is “0.20”.

  In this case, in step S3 shown in FIG. 3, the horizontal line correlation Hs is calculated as the feature amount of the code portion of the one-dimensional color code included in the extracted block. In this case, the color plane selection in step S4 employs a method of selecting the color plane having the largest feature amount. For example, the color plane having the highest preset priority order is selected from among the R, G, B planes 41R, 41G, 41B having the horizontal line correlation Hs of the extracted block pixel data equal to or greater than a predetermined value. Also good. When a color plane is selected in this way, even if a red character or the like is written on a black one-dimensional code, the R plane 41R is selected and used for the recognition processing in step S5. It is possible to perform highly accurate recognition processing without being affected. However, in this case, the R, G, and B planes 41R, 41G, and 41B are preset in descending order of priority, and the horizontal line correlation Hs of the pixel data of the extracted block is, for example, “0.85” or more. It is assumed that the color plane having the highest priority among the color planes is selected. In this way, the color plane selection criterion in step S4 may be selected in accordance with the state of the color code and the processing performed in the recognition processing in step S5.

  FIG. 12 is a block diagram for explaining another embodiment of the code recognition apparatus according to the present invention. In the above embodiment, the code recognition device is configured by software. However, in this embodiment, the code recognition device is configured by hardware.

  12 includes an image reading unit 61, an image data generation unit 62, a code determination unit 63, a color plane selection unit 64, a code recognition processing unit 65, and a recognition result output unit 66.

  The image reading unit 61 includes a digital camera or the like, and corresponds to the image reading device 9 shown in FIG. The image data generation unit 62 performs the processes of steps S1 and S2 shown in FIG. The code determination unit 63 performs the process of step S3 shown in FIG. The color plane selection unit 64 performs the process of step S4 shown in FIG. The code recognition processing unit 65 performs the process of step S5 shown in FIG. The recognition result output unit 66 constitutes an interface for outputting the recognition result obtained by the code recognition processing unit 65 to the outside of the code recognition device 51.

  Needless to say, the processing performed by three or less of the image data generation unit 62, the code determination unit 63, the color plane selection unit 64, and the code recognition processing unit 65 may be realized by software processing.

  As described above, in the present invention, after a code is read and a color plane for each of a plurality of color components is created from input color image data, the image data of each color plane is divided into units of rectangular blocks, and each block is divided. The feature amount of the code is calculated, and the presence / absence and type of the code are determined from the feature amount. The feature amount of the code for each color plane is compared, and for example, the presence / absence and type of the code is determined using the color plane in which the feature amount is most strongly expressed, and the code is recognized. As a result, even when a code of a color other than black is processed, it is possible to perform processing by selecting a color plane suitable for code discrimination and recognition, so that the code recognition process can be performed accurately and at high speed. Can be done.

  In order to perform code recognition processing with high accuracy and high speed, a color plane having a block in which the variation of the pixel value in the block is maximum within a predetermined range is selected, or white or black pixels in the block are selected. Select a color plane with a block whose ratio falls within a predetermined range, select a color plane with a block whose number of edges in the horizontal and vertical lines within the block falls within a predetermined range, or identify within a block A color plane having a block in which the correlation between the horizontal lines and the correlation between the vertical lines falls within a predetermined range can be selected.

  In this way, since there are a plurality of feature quantities different from the background part in the code part of the one-dimensional code and the two-dimensional code, the code is unified by selecting a color plane by combining a plurality of these feature quantities. And code recognition can be performed with higher accuracy.

  As described above, according to the present invention, it is possible to automatically select a color plane suitable for the color of the read code and perform a recognition process. But it can be recognized reliably.

  The present invention can be applied to determination of whether a code is a unitary code or a two-dimensional code and code recognition regardless of the color of a code to be read.

  While the present invention has been described with reference to the embodiments, it is needless to say that the present invention is not limited to the above-described embodiments, and various modifications and improvements can be made within the scope of the present invention.

Claims (11)

Dividing the color image data of the read code into color components of a plurality of color planes;
An extraction step of extracting a rectangular region from an arbitrary position of each color plane;
A calculation step for calculating a feature amount of a code included in a rectangular area of each color plane;
A selection step of selecting one color plane based on the feature amount ;
Look contains a recognition step of recognizing code based on the color component data of the selected color plane,
The calculating step calculates a variation in pixel values in the rectangular area as a feature amount, and the selecting step selects a color plane having a rectangular area where the variation is maximum within a predetermined range. A code recognition method.   The calculation step calculates the minimum value of the pixel value in the rectangular area as a feature amount, and the selection step selects a color plane having a rectangular area where the minimum value is the minimum within a predetermined range. The code recognition method according to claim 1, wherein the code recognition method is characterized.   The calculating step calculates a ratio between white pixels and black pixels in the rectangular area as a feature amount, and the selecting step is within a predetermined range including the reference value, and is the largest in the reference value. 3. The code recognition method according to claim 1, wherein a plane having a block having a close value is selected.   The calculating step calculates the number of pixel value change points in the horizontal line and the vertical line in the rectangular area as a feature amount, and the selecting step includes the number within a predetermined range and a maximum value. The code recognition method according to claim 1, wherein a plane having a block is selected.   The calculating step calculates a correlation between specific horizontal lines and vertical lines in the rectangular area as a feature amount, and the selecting step selects a block in which the correlation is within a predetermined range and becomes the maximum. The code recognition method according to any one of claims 1 to 4, wherein a plane to be selected is selected. A computer, characterized in that it comprises a procedure for executing the steps of the code recognition method according to any one of claims 1-5, program. Means for dividing the color image data of the read code into color components of a plurality of color planes;
Extraction means for extracting a rectangular region from an arbitrary position of each color plane;
A calculation means for calculating a feature amount of a code included in a rectangular area of each color plane;
Selection means for selecting one color plane based on the feature amount ;
Recognizing means for recognizing the code based on the color component data of the selected color plane ,
The calculating means calculates a variation in pixel values in the rectangular area as a feature amount, and the selecting means selects a color plane having a rectangular area where the variation is maximum within a predetermined range. Electronic devices.   The calculation means calculates the minimum value of the pixel values in the rectangular area as a feature amount, and the selection means selects a color plane having a rectangular area where the minimum value is minimum within a predetermined range. The electronic device according to claim 7, wherein the electronic device is characterized.   The calculating means calculates a ratio between white pixels and black pixels in the rectangular area as a feature amount, and the selecting means is within a predetermined range including the reference value and closest to the reference value. 9. The electronic device according to claim 7, wherein a plane having a block to be a value is selected.   The calculating means calculates the number of change points of the pixel values in the horizontal line and the vertical line in the rectangular area as a feature amount, and the selecting means is such that the number is within a predetermined range and is the maximum. The electronic device according to claim 7, wherein a plane having a block is selected.   The calculating means calculates a correlation between specific horizontal lines and vertical lines in the rectangular area as a feature amount, and the selecting means calculates a block in which the correlation is within a predetermined range and becomes the maximum. The electronic device according to claim 7, wherein the plane is selected.

Images