JP5683891B2 - Stacked bar code reader and stack bar code reading method - Google Patents

Description

  The present invention relates to a stack bar code reader and a stack bar code reading method for optically reading a stack bar code.

  Conventionally, barcodes in which character information is encoded into barcodes and displayed as bars and spaces are widely used as a means for assigning unique information to media such as paper and plastic and automatically recognizing them. doing. For example, a one-dimensional barcode is used for general consumer goods. In recent years, demands for increasing the amount of information on barcodes have increased, and two-dimensional barcodes that have a much higher capacity (information density) than one-dimensional barcodes have appeared. And readers corresponding to this two-dimensional bar code are gradually spreading.

  Among these two-dimensional barcodes, there is a method of increasing the amount of information by stacking one-dimensional barcodes, which is called a stack barcode. A representative example is a code-based stack barcode called PDF417. This code system PDF417 is internationally standardized as ISO / IEC15438.

  As a reading device for reading such a stack type barcode, for example, there is a symbol information reading device disclosed in Patent Document 1. The symbol information reading apparatus disclosed in Patent Document 1 captures a stack barcode with a two-dimensional imaging device, expresses the quality state of the stack barcode numerically, and can grasp the deterioration state of the stack barcode. It is a thing.

JP-A-8-241369

  However, when the resolution of the imaging optical system that captures the stack barcode does not reach a certain level, the scanning reflectance waveform of the stack barcode varies. For this reason, the symbol information reading apparatus disclosed in Patent Document 1 also has a problem that the boundary between the bar and the space cannot be accurately detected due to fluctuations in the scanning reflectance waveform. As a result, the reading accuracy of the stack barcode may be lowered.

  In view of such circumstances, the present invention is not easily affected by fluctuations in the scanning reflectance waveform of a stack barcode, and a stack barcode reader and stack type capable of accurately reading the stack barcode. An object is to provide a barcode reading method.

In order to achieve such an object, a stack type barcode reader according to the present invention includes an imaging unit that captures a stack type barcode in which character information is encoded into a barcode and displayed as a bar and a space, and the imaging unit. An image memory for storing the image data of the stack type barcode imaged by the data processing unit, and a data processing unit having a decoding unit for decoding the barcode of the stack type barcode into character information based on the image data, The decoding unit constitutes the image data, threshold calculating means for calculating a threshold for measuring the line width between the bar and space of the stack barcode based on the luminance values of a plurality of pixels constituting the image data. An n-th order differential calculation method for differentiating a scanning reflectance waveform indicating the luminance values of a plurality of pixels to calculate an n-th order differential value (n is 1 or 2). A line width measuring means for measuring the line width of the bar and space of the stacked bar code based on the threshold value or the nth-order differential value, and the bar code of the stack bar code measured by the line width measuring means. And a stack type barcode reader having a T sequence measuring means for measuring a T sequence from the line width of the space, and a code word acquiring means for acquiring a code word corresponding to the T sequence measured by the T sequence measuring means. The T sequence measurement means determines the validity of the T sequence measured based on one of the threshold value and the nth-order differential value, and when the T sequence is not valid, reading stacked bar code determining the efficacy of T sequences that are measured based on the other of the n-th order differential value It characterized in that it was the location.

  Furthermore, in the stack type barcode reader according to the present invention, the n-th order differential value is a second-order differential value, and a zero-cross point of the second-order differential value is used as a boundary between the bar and the space. To do.

  In the stacked barcode reader according to the present invention, the threshold value may be a fixed threshold value based on an average value of the image data, and a maximum peak value and a minimum peak value of the bar and the space adjacent to each other in the image data. Or a local threshold value based on the average value.

Further, the stack type barcode reading method according to the present invention is based on image data obtained by imaging a stack type barcode in which character information is encoded into a barcode and displayed in a bar and a space. Threshold value for calculating a threshold value for measuring the line width between the bar and space of the stack barcode according to the luminance values of a plurality of pixels constituting the image data when decoding the barcode barcode into character information A calculation step, an n-order differential calculation step of calculating an n-th order differential value (n is 1 or 2) by differentiating a scanning reflectance waveform indicating luminance values of a plurality of pixels constituting the image data, the threshold value or A line width measuring step for measuring the line width of the bar and space of the stacked barcode based on the nth-order differential value, and the line width measuring step A T-sequence measurement step of measuring a T-sequence from the line width of the bar and space of the stacked barcode, and a code-word acquisition step of acquiring a code word corresponding to the T-sequence measured in the T-sequence measurement step A stack type bar code reading method including : determining the validity of a T sequence measured based on one of the threshold value and the n-th order differential value in the T sequence measurement step; If not, a stack type barcode reading method for determining the validity of the T sequence measured based on the other of the threshold value and the nth-order differential value is provided.

  The stacked barcode reading method according to the present invention further determines the validity of the T sequence based on whether or not the cluster number calculated from the T sequence measured in the T sequence measurement step is equal to the theoretical value. It is characterized by that.

According to the stack type barcode reader and the stack type barcode reading method according to the present invention, two types of low level decoding methods, ie, a threshold method and an nth-order differentiation method are provided, and the resolution of the imaging optical system is varied. By properly using these, it is difficult to be affected by fluctuations in the scanning reflectance waveform of the stack barcode, and the stack barcode can be read with high accuracy.

It is a block diagram which shows the electric constitution of the stack | stuck type barcode reader which concerns on Embodiment 1 of this invention. It is a block diagram of the low level decoding part of the stack type barcode reader according to the first embodiment. It is a figure which shows an example of a stack | stuck type | mold barcode. It is the enlarged view which expanded the part of the big bar in a start pattern among stack-type barcodes, and its periphery. It is a flowchart which shows the flow of the stack | stuck type | mold barcode reading method which concerns on Embodiment 1 of this invention. It is a flowchart which shows the flow of a stack | stuck type | mold barcode position detection in the flowchart shown in FIG. It is a flowchart which shows the flow of the structural analysis (column analysis and row analysis) in the flowchart shown in FIG. It is the graph calculated by giving a predetermined process to correction | amendment image data (refer FIG. 3). It is a wave form diagram which shows the pixel value of each line in the horizontal direction of correction | amendment image data. It is explanatory drawing for demonstrating a mode that the division | segmentation of the row | line | column of a stack type barcode is specified. 6 is a flowchart showing a flow of low-level decoding in the flowchart shown in FIG. 5. It is a wave form diagram which shows a part of scanning reflectance waveform which shows the luminance value of each pixel which comprises the row | line | column of a stack | stuck type | mold barcode. It is a flowchart which shows the flow of a X sequence measurement process. It is the figure which extracted a part of scanning reflectance waveform which shows the luminance value of the row | line | column of a stack | stuck type | mold barcode. It is a figure for demonstrating an nth-order differentiation system. It is a flowchart which shows the flow of T sequence measurement processing. FIG. 7 is a diagram showing a part of a scanning reflectance waveform that indicates a luminance value of a row of a stack barcode in the stack barcode reading method according to a modification of Embodiment 1 of the present invention.

Embodiments of the present invention will be described below.
Embodiment 1 of the Invention

1 to 17 show a first embodiment of the present invention.
(Stacked barcode reader)

  FIG. 1 is a block diagram showing an electrical configuration of a stack type barcode reader 1 according to Embodiment 1 of the present invention. FIG. 2 is a block diagram of the low-level decoding unit of the stacked barcode reader according to the first embodiment.

  In FIG. 1, the stack barcode reader 1 includes an imaging device 11 having a contact image scanner 11 a and a card transport mechanism 11 b, an image memory 12, and a data processing device 13. The data processing device 13 includes a position detection unit 13a, an inclination correction unit 13b, a structure analysis unit 13c, and a decoding unit 13d. Further, a stack type barcode 21 which is a kind of two-dimensional barcode is printed on a record carrier 2 such as a card. The decoding unit 13d includes a low level decoding unit 13d1 and a high level decoding unit 13d2.

  The image scanner 11a of the imaging device 11 obtains image data by converting the reflected light of the light irradiated on the stack barcode 21 printed on the record carrier 2 into an electrical signal by photoelectric conversion. The obtained image data of the stack barcode 21 is stored in the image memory 12. The data processing device 13 captures the image data of the stack barcode 21 from the image memory 12 and performs various processes on the captured image data.

  As illustrated in FIG. 2, the low-level decoding unit 13 d 1 includes an X sequence measurement unit 14, a T sequence measurement unit 15, and a codeword conversion unit 16.

  The X sequence measurement unit 14 includes a threshold value calculation unit 14a and a line width measurement unit 14b. The threshold value calculation means 14a is composed of a fixed threshold value calculation means 14a1, a local threshold value calculation means 14a2, and an n-th order differential calculation means 14a3.

  The T sequence measurement unit 15 includes a sequence conversion unit 15a, a line width normalization unit 15b, and a determination unit 15c. The sequence conversion means 15a has a function of converting the line width array calculated by the X sequence into a line width array for the T sequence. The determination unit 15c has a function of comparing a cluster number, which will be described later, with a theoretical value.

The code word conversion unit 16 includes a table memory 16a, and a code word conversion table indicating a correspondence relationship between T sequences and code words is stored in the table memory 16a.
(Structure of stack barcode PDF417)

  FIG. 3 is a diagram showing an example of the stack barcode 21 (stack barcode with the label structure of PDF417).

  In FIG. 3, the stack type barcode 21 has a vertical direction (Y direction) in the drawing in the row direction (a direction in which a plurality of rows are arranged) and a horizontal direction (X direction) in the column direction (a plurality of columns are arranged in the direction). In the column direction, and is roughly divided into five. That is, as shown in FIG. 3, it is divided into five columns, the data column is arranged in the center, the left and right indicators on the left and right of the data column, and the start pattern and stop on the left and right, respectively. A pattern is placed. Each bar code is composed of a black bar and a white space, and the start pattern and stop pattern are each formed with a wide bar called a big bar. Yes. The big bar in FIG. 3 is the widest bar in the black part on the left side for both the start pattern and the stop pattern.

  Further, as shown in FIG. 3, each of the data column, the left row indicator, and the right row indicator is composed of six rows (row 1 to row 6) in the row direction. Each of the three columns (data column, left row indicator, right row indicator) constituting each row includes one code word. For example, in the data column of the stack barcode 21, six code words are composed of one (for one data column) × 6 rows. The code word is a basic unit for encoding a value representing a certain number, character or other symbol, or a value associated therewith.

  Each column except the stop pattern has four bars and four spaces, and each line width is an integral multiple of the minimum width called a module (unit). The line width of one code word is 17 modules, and the maximum unit of each bar and space is 6. The stop pattern has five bars and four spaces, and the line width of one code word is 18 modules.

  Note that the stack barcode 21 shown in FIG. 3 has one data column, but the present invention is not limited to this, and a plurality of data columns may be used.

The stack bar code reading method according to the first embodiment of the present invention will be described below.
(Stacked barcode reading method)

  FIG. 5 is a flowchart showing the flow of the stack barcode reading method according to the first embodiment. Based on FIG. 5, the flow of the stack barcode reading method according to the first embodiment will be outlined.

First, the position of the stack barcode 21 is detected (step S31). Specifically, as shown in FIG. 1, the image scanner 11a of the stack barcode reader 1 irradiates the stack barcode 21 of the record carrier 2 that has moved to a predetermined position along the card transport mechanism 11b. The reflected light is converted into an electrical signal by photoelectric conversion to obtain image data. Then, the obtained image data of the stack barcode 21 is stored in the image memory 12. The image memory 12 is composed of a finite number of pixels arranged two-dimensionally.
In the first embodiment, the captured image data of the stack barcode 21 represents the luminance value (which is a pixel value) of each pixel by a numerical value. This luminance value is expressed as an integer value from 0 to 255, for example. The image memory 12 may be any memory such as RAM, SDRAM, DDRSDRAM, and RDRAM as long as it can store image data.

  Thereafter, the position detection unit 13a of the data processing device 13 reads the image data stored in the image memory 12, and detects the position of the stack barcode 21 in the image data.

  The stack barcode 21 whose position has been detected in this manner is cut out by the data processing device 13 (step S32), and the cut out image data is stored in the image memory 12.

  Next, the data processor 13 corrects the inclination of the stack barcode 21 (step S33). More specifically, the inclination correction unit 13b of the data processing device 13 receives the image data cut out in step S32, that is, the image data of the stack barcode 21 captured in an inclined state as shown in FIG. The image data is read and converted into corrected image data with a tilt angle of zero as shown in FIG.

  FIG. 4 is an enlarged view of the stack bar code 21 in which the portion of the big bar in the start pattern and its periphery are enlarged. FIG. 4 is a diagram illustrating a state where the stack barcode 21 is imaged in an inclined state by the image scanner 11a of the imaging device 11. When the stack type barcode 21 is scanned by the imaging device 11, the row direction of the stack type barcode 21 is not necessarily positioned perpendicular to the card transport mechanism 11b. The image data read by the image scanner 11a may be tilted as shown in FIG. 4, for example. In other words, the row direction does not match the vertical direction of the image data stored in the image memory 12. In such a case, along with the coordinate conversion, the reliability of decoding has been impaired due to noise caused by the quantization error that occurs due to the conversion of image data between pixels. Therefore, in the first embodiment, interpolation processing and averaging processing for smoothing are performed on the image after coordinate conversion to reduce the influence of such noise.

  Next, structural analysis (column analysis and row analysis) is performed in the data processing device 13 (steps S34 and S35). More specifically, the structure analysis unit 13c of the data processing device 13 analyzes the structure of the stack barcode 21 based on the corrected image data converted in step S33. In the structure analysis unit 13c, column analysis (step S34) and row analysis (step S35) are performed.

  Finally, decryption processing is performed in the decryption unit 13d of the data processing device 13 (steps S36 and S37). More specifically, the decoding unit 13d of the data processing device 13 performs the decoding process of the stack barcode 21 based on the structure of the stack barcode 21 analyzed in steps S34 and S35. At this time, two types of decoding processes, a low-level decoding process and a high-level decoding process, are performed in PDF417, which is a type of stack barcode 21, and a low-level decoding process is performed in the low-level decoding unit 13d1 of the decoding unit 13d. (Step S36), the high level decoding unit 13d2 performs a high level decoding process (Step S37). In the low-level decoding process, the code word composed of the line width of the bar and the line width of the space constituting the data column is once converted into intermediate information indicated by 929 codes with numbers 0 to 928. It is processing. Next, in the high-level decoding process, the information is once converted into intermediate information represented by 929 codes and then decoded (decoded) into the final language based on a predetermined rule. As the final language, PDF417 includes ASCII characters, binary expressions, and other characters.

Hereafter, each process (step S31-step S37) of the flowchart shown in FIG. 5 is each explained in full detail.
(Stacked bar code position detection)

  FIG. 6 is a flowchart showing the flow of stack bar code position detection (step S31) in the flowchart shown in FIG. FIG. 4 is an enlarged view of the stack bar code 21 in which the portion of the big bar in the start pattern and its periphery are enlarged.

  In FIG. 6, first, a variable i (i is an integer) is initialized by substituting 0, and then returns to a predetermined origin (step S61). Specifically, FIG. 4 shows an area surrounding a part of the stack barcode 21 in the image data stored in the image memory 12, and the pixel S is a predetermined origin position in the upper left. Therefore, in step S61, the row of one pixel in the horizontal direction (X direction) is set as the horizontal line HL (0), and the image data of the first pixel S is read.

  Next, for the i-th pixel column in the vertical direction (Y direction) from the horizontal line HL (0), image data is read for each pixel in the X direction on the horizontal line HL (i) at Y = i. A point where the image changes from “bright” to “dark” (referred to as point P (i)) is searched. Specifically, image data of each pixel, that is, a pixel (hereinafter referred to as a “space”) that indicates “bright” (white portion) whose luminance value is higher than a predetermined threshold, and “ It is divided into pixels (hereinafter referred to as “bars”) indicating “dark” (black portions), and the positions of the pixels that first changed from space to bar are detected. The changed pixel position becomes the boundary between the space and the bar.

  Further, on the same horizontal line HL (i), a point where the image first changes from “dark” to “bright” (referred to as point Q (i)) is continuously searched. Specifically, the position of the pixel changed from a bar (a pixel indicating “dark”) to a space (a pixel indicating “bright”) is detected (step S62).

  Next, it is determined whether or not i = r−1 (r: the number of rows (number of pixels) in the Y direction in FIG. 4) is satisfied (step S63). If it is determined that i is not satisfied, i is incremented by 1 (step S64), and the process of step S62 is performed again. On the other hand, when the search starts from the pixel S shown in FIG. 4 and the luminance value of each pixel is searched in the X direction and the search is completed up to the last pixel of the last row, it is determined that i = r−1 is satisfied. To do. Thus, when it determines with satisfy | filling, a process is moved to following step S65. At this stage, a set of r points (P (i), Q (i)) (i = 0, 1,..., R−2, r−1) is obtained.

  Next, the coordinates of point A and point C in FIG. 4 are obtained (step S65). More specifically, the positions of the pixels corresponding to the points A and C are detected. In the X direction, d (i) = | P (i) −Q (i) | is calculated, and the minimum value min (d (i)) of d (i) is obtained. This min (d (i)) corresponds to two points of the big bar point A and point C where the distance between the PQs is the shortest (see FIG. 4). Thereby, the Y coordinates (positions of the corresponding pixels in the Y direction) of the points A and C are Ay = max (i) and Cy = min (i), respectively. In addition, since the X coordinates of the points A and C (the positions of the corresponding pixels in the X direction) are the X coordinates of the point P with respect to each i, the X coordinates of Ax = P (i = Ay) and Cx = This is the X coordinate of P (i = Cy).

  Next, the coordinates of point B and point D in FIG. 4 are obtained (step S66). More specifically, the range of i in which d (i) is constant is specified. The Y coordinates of the big bar points B and D are min (i) and max (i) in the range. Further, since the X coordinates of the points B and D are the X coordinates of the points P and Q with respect to the respective i, the X coordinates of Bx = P (i = By) and Dx = Q (i = Dy) respectively. X coordinate.

  In this way, through the processing of step S61 to step S66, the four corners of the big bar, that is, the coordinates of the points A, B, C, and D (the positions of the corresponding pixels) are determined. Can be calculated by atan ((Ay−Dy) / (Ax−Dx)).

  In order to increase the accuracy of inclination detection, the same operation is performed by exchanging the roles of X and Y in FIG. 4, and the four corners of the big bar, that is, point A ′, point B ′, point C ′, point The coordinates of D ′ are obtained, and the average of the coordinates of the points A, B, C, D may be taken as the final coordinate value.

In FIG. 6, the coordinates of the points A and C are obtained (step S65), and the coordinates of the points B and D are also obtained (step S66). However, the present invention is not limited to this. The inclination may be detected by obtaining only the coordinates of the points A and C.
(Cutting stack barcode)

When the coordinates of the four corners of the big bar are determined through the processing of steps S61 to S66, the stack barcode 21 is cut out (step S32). Note that the cutting is performed in a rectangular shape that is slightly larger than the entire stack barcode 21.
(Stacked bar code tilt correction)

  Next, the inclination correction of the stack barcode 21 is performed (step S33).

  FIG. 3 shows a state in which the image data of the stack barcode 21 tilted at the above-described angle (= atan ((Ay−Dy) / (Ax−Dx))) is converted into corrected image data having a tilt angle of zero. FIG. In other words, the image data of the stack barcode 21 shown in FIG. 3 has the column direction and the row direction corresponding to the horizontal direction (X direction) and the vertical direction (Y direction) of the image memory 12 area shown in FIG. It is corrected as follows.

  In FIG. 3, conversion to corrected image data with a tilt angle of zero is performed using the tilt angle of the stack barcode 21. For example, since it can be performed by a known method such as affine transformation, a detailed description thereof is omitted here.

In the stacked barcode reading method according to the first embodiment, in order to alleviate the influence of the quantization error caused by the coordinate transformation, the interpolation processing or averaging is performed on the image after the coordinate transformation for the purpose of smoothing. Process. Specifically, as shown in FIG. 9B and FIG. 12, the luminance value of each pixel is set to a luminance value indicating a space (for example, 100) and a luminance value indicating a bar (for example, 0). The corrected image data is smoothed by applying various filters such as a median filter, an edge preserving filter, an adaptive winner filter, and a moving average filter.
(Structural analysis)

  FIG. 7 is a flowchart showing the flow of structural analysis (steps S34 and S35) in the flowchart shown in FIG.

  In FIG. 7, first, horizontal projection processing is performed (step S91). Specifically, based on the corrected image data of the stack barcode 21 shown in FIG. 3, for each pixel arranged in the horizontal direction, the sum of the luminance values of all the pixels arranged in each vertical direction is calculated, This is a process of storing in the image memory 12.

  Thereafter, with respect to the sum of the obtained luminance values, a difference is taken between pixels adjacent in the horizontal direction, and the start pattern and the stop pattern are analyzed using the difference value (step S92).

  As shown in FIG. 3 and the like, the start pattern and stop pattern are invariant to all rows (6 rows in the first embodiment), unlike the code words of other data columns, left row indicators, and right row indicators. Therefore, by taking a horizontal projection, the start pattern or stop pattern is averaged, and it is difficult to be affected by uneven brightness and dirt, and stable detection is possible.

  FIG. 8 is a diagram obtained by performing predetermined processing on the corrected image data (see FIG. 3). More specifically, FIG. 8A shows the projection of the corrected image data (see FIG. 3) calculated for each pixel in the vertical direction. In FIG. It is a distribution indicating the sum of luminance values of all pixels in the direction. FIG. 8B shows the difference calculated for the graph of FIG. Specifically, the difference of the sum of the obtained luminance values is calculated between the pixels adjacent in the horizontal direction of the image data stored in the image memory 12, and the distribution of the difference values is represented. Thereby, the position of the pixel which changed from the bar (or space) to the space (or bar) on the horizontal line L (i) can be detected.

  The start pattern is detected by measuring the interval (run length) between the points where the bar and the space change (hereinafter referred to as “line width”) using the graph shown in FIG. 8A or FIG. 8B. . In the present embodiment, this is done by counting (counting) the number of pixels in the horizontal direction. For example, when the line width of each bar of the start pattern and the line width of each space are measured, a line width array as shown in StartWork = [47, 6, 5, 6, 7, 6, 6, 18] is obtained. On the other hand, the ratio between the length of the bar and the space of the start pattern is a line width array as shown in StartMark = [8, 1, 1, 1, 1, 1, 1, 3] according to the PDF417 standard. Yes. Here, for example, a normalized correlation R can be used as an index for examining the similarity between the two. Considering the normalized correlation R in FIG. 8B, R = corrcoef (StartWork, StartMark) = 0.993, and R is sufficiently close to 1. Therefore, in this case, it is determined that the start pattern is appropriately detected. Note that the description of the normalized correlation R of the stop pattern is the same as that of the start pattern, and thus the description thereof is omitted here.

  In the first embodiment, the normalized correlation R is used as a measure of similarity. However, the present invention is not limited to this, and for example, a sum of absolute differences, a sum of products, or the like may be used as appropriate.

  Thus, after it is determined whether or not the start / stop pattern is properly detected (step S93), if it is determined that the start / stop pattern is properly detected, the process proceeds to step S95. . On the other hand, if it is determined that the start / stop pattern is not properly detected, the process is terminated because the decoding is impossible, that is, the decoding is impossible (step S94).

  Next, peak detection is performed on the graph of FIG. 8B (step S95). More specifically, as shown in FIG. 3, in the start pattern, the left row indicator, the data column, the right row indicator, and the stop pattern, transition from space to bar is performed in all rows at the boundary of each column. Therefore, a large difference value appears in this portion on the distribution of difference values in horizontal projection (upward arrow in FIG. 8B). Then, in order to obtain the difference value (peak value) indicated by the upward arrow in FIG. 8B, an appropriate threshold value is prepared, and the presence or absence of a peak can be determined based on whether or not the threshold value is exceeded. . If the peak is determined, the pixel memory having the peak value is stored in the image memory 12.

  Next, column boundary detection is performed (step S96). More specifically, the boundaries of the five columns of the start pattern, the left row indicator, the data column, and the right row indicator stop pattern are detected based on the position of the pixel having the peak value stored in step S95. The interval between pixels having adjacent peak values represents the column width, and is stored in the image memory 12 in the same manner. Here, when there is a stain in the vicinity of this boundary in a specific row, there may be a case where the change point indicating the boundary of each column cannot be correctly detected during the scanning of each row, but as in the first embodiment, By using the horizontal projection, the influence of this dirt is reduced by the averaging operation, and the adverse effect on the detection of the boundary of each column can be suppressed.

  Next, a row boundary is detected (step S97). More specifically, a description will be given with reference to FIGS. 9 and 10. FIG. 9 is a waveform diagram showing the luminance value of each line L in the horizontal direction of the averaged corrected image data. FIG. 10 is an explanatory diagram for explaining a state in which each row separator of the stack barcode 21 is specified. In the first embodiment, one line L is a matrix of 1 row × n columns. Here, n is the width of the evaluation range.

  In FIG. 9A, the detection of a row boundary is performed by first setting a set of a plurality of continuous lines L on the corrected image data. For example, a set (line set S1) composed of three continuous lines L (1), L (2) and L (3) (hereinafter, abbreviated as “L1, L2 and L3”) is set (see FIG. 10). ). In the first embodiment, three continuous lines L are set, but the number is not limited to three.

  When two arbitrary lines are selected for these three lines L and their normalized correlations are obtained, a total of three correlation values (L1 and L2, L1 and L3, L2 and L3) are obtained. If the smallest value of these three correlation values is larger than a predetermined threshold value r0, the waveforms of L1, L2, and L3 (indicating the luminance value of each pixel in line L) are correlated with each other. 1 is given to this line set S1 (see FIG. 10). On the other hand, if it is smaller than the predetermined threshold value r0, it is determined that the waveforms of the two lines L from which the smallest value is obtained are at least uncorrelated, and 0 is given to this line set S1.

  For example, for a set of lines L1, L2 and L3 formed in the horizontal direction, as shown in FIG. 9B, the lines L1, L2 and L3 are all included in the same row 1 (the uppermost row). Therefore, the smallest value among the three correlation values is larger than the threshold value r0 and 1 is given (in FIG. 10, the determination value of the line set S1 is 1).

  Next, the above-mentioned set is moved relative to the corrected image data in the row direction (vertical direction) (for example, considering the line set S2 including L2, L3, and L4), and a correlation value sequence is calculated. By determining whether or not the calculated minimum value of the correlation value sequence is larger than the threshold value r0, a determination value of 0 or 1 is given to the line set.

  Similarly, when the above-described processing (processing for giving a judgment value) is repeated up to lines Ln-2, Ln-1, and Ln (line set Sn-2), n-2 1s and 0s are formed in total. A series is obtained (right column in FIG. 10).

  Here, when three lines L are selected from the same row, the line set consisting of these three lines L has a value close to 1, and the three lines L are selected across adjacent rows. In this case, the line set including these three lines L has a relatively small value. More specifically, in FIG. 9B, the line L above the line L18 belongs to the row 1, the line L below the line L18 belongs to the row 2, and the line L18 includes the row 1 and the row 2 The line L is located at the boundary.

  Therefore, in the line sets S16 to S18 including the line L18, the determination value is almost 0, and in the other line sets, the determination value is 1 (see FIG. 10). A line set whose judgment value is close to 0 represents a row switching position. By scanning a series of judgment values and identifying a place where the judgment value is close to 0, the row switching position is known. Can do. In the first embodiment, since the similarity between the lines L is evaluated by a correlation function, as shown in FIG. 9A, a common pattern (bar In addition, by making the range excluding the same line width of the space as the evaluation range, useless computation (calculation cost) can be omitted.

  As described above, step S97 is performed by the flow of setting a line group → calculating a correlation value sequence → comparing the minimum value of the correlation value sequence with a threshold value → giving a determination value of 0 or 1 → scanning a sequence of determination values. The row boundary is detected. In the first embodiment, the process is performed from row 1 to the last line L of row 6.

  By detecting the boundary of the row, it is determined to which row each line L of the corrected image data belongs, and the processing is ended (decoding is possible) (step S98).

In the first embodiment, the local noise of the line L can be reduced by calculating the average of the luminance values of the points of the plurality of lines L included in each row. Further, the average luminance value of the pixel is used as a representative of the characteristics of the row, and can be used for a decoding process described later. Thereby, decoding (decoding) accuracy can be improved.
(Low level decoding)

  FIG. 11 is a flowchart showing the flow of low-level decoding (step S36) in the flowchart shown in FIG.

In this low level decoding, as shown in FIG. 11, with respect to the scanning reflection characteristics of each row of the stack barcode 21, segments having the column boundaries determined by the column analysis at both ends (that is, scanning reflectance waveforms) are obtained. Take out an X sequence measurement to obtain a numerical sequence called an X sequence formed by the arrangement of the line width of the bars and the space width of one column, and two adjacent bars and spaces in the X sequence T sequence measurement for obtaining a numerical sequence called a T sequence by adding them while shifting, and code word conversion for determining a code word corresponding to each T sequence with reference to a code word conversion table prepared in advance Become.
(X sequence measurement in low-level decoding)

  In PDF417, which is a kind of the stack barcode 21, two decoding processes, a low-level decoding process and a high-level decoding process, are performed. The code word consisting of the line width of the bar and the line width of the space constituting the data column consisting of the width is converted once into intermediate information indicated by 929 codes, numbered from 0 to 928. is there. Next, in the high-level decoding process, the information is once converted into intermediate information represented by 929 codes and then decoded (decoded) into the final language based on a predetermined rule. As the final language, PDF417 includes ASCII characters, binary expressions, and other characters.

  In FIG. 11, first, the X sequence measurement unit 14 of the low level decoding unit 13d1 performs an X sequence measurement process (step S131). In a general X sequence measurement process, the line width measurement unit 14b of the X sequence measurement unit 14 scans a line waveform obtained by averaging a plurality of lines L included in a row in each row, thereby obtaining the line width of the bar. After the line width measurement for obtaining the line width of the space is performed, the line width expressed by the number of pixels is converted into the module number expression to obtain an X sequence. Note that the scanning line is, for example, a plurality of lines L formed in the middle of each row in each row, that is, a row example of pixels formed of a plurality of rows. Furthermore, the average of the luminance values of the pixels constituting the plurality of lines L is calculated, and the averaged luminance value of the pixels is used in the decoding process as a scanning reflectance waveform indicating the luminance value of the pixels in the row.

  FIG. 12 is a waveform diagram showing a part of a scanning reflectance waveform indicating a luminance value of a pixel obtained by scanning a line waveform obtained by averaging a plurality of lines L included in a row in a certain row. The line width of the bar and the line width of the space are obtained by counting the number of pixels between the changing points of the scanning reflectance waveform indicating the luminance value of the pixels. Here, the change point indicates a point that intersects a predetermined threshold (not shown) on the scanning reflectance waveform indicating the luminance value of the pixel. As shown in FIG. 3 and the like, the PDF 417, which is a kind of code system of the stack barcode 21, is composed of five columns: a start pattern, a left row indicator, a data column, a right row indicator, and a stop pattern. Further, since each column except the stop pattern is composed of four bars and four spaces, if there is no damage to the scanning reflectance waveform indicating the luminance value of the pixel, a total of eight line widths are obtained. can get.

On the other hand, when there is an abnormality in the scanning reflectance waveform indicating the luminance value of the pixel, a total of 8 or more (9) line widths are obtained. In such a case, it is found that the column is damaged, and a flag indicating that the reliability of this column is low is set. When this operation is executed for all the columns included in a certain row, the same processing is executed for the next row. When the line widths are obtained for all the rows, the X sequence measurement process in step S131 (see FIG. 11) is terminated.
(T-sequence measurement in low-level decoding)

  Next, the T sequence measurement unit 15 of the low level decoding unit 13d1 performs T sequence measurement processing (step S132).

  Specifically, the sequence conversion means 15a of the T sequence measurement unit 15 converts the X sequence into a T sequence. This is a sum of two adjacent Xi and Xi + 1. When the X sequence of the left column (data column) shown in FIG. 12 is converted into a T sequence, [6 (5 + 1), 2 (1 + 1), 2 (1 + 1), 2 (1 + 1), 2 (1 + 1), 6 (1 + 5), 7 (5 + 2)].

  Next, the line width normalization unit 15b of the T sequence measurement unit 15 performs line width normalization processing. This line width normalization process refers to a process of converting the line width expressed by the number of pixels representing the measured line width of the bar and the line width of the space into the module number expression. One module corresponds to the minimum width of the bar or space. Specifically, in PDF417, which is a kind of stack barcode 21, one column including a data column is composed of 17 modules. If the line width in one column is W1, W2, W3, W4, W5, W6, W7, W8, and Wc = W1 + W2 + W3 + W4 + W5 + W6 + W7 + W8, the normalized line width is Xi = Wi * 17 / Wc (i = 1, 2 ,..., 7, 8). For example, the X sequence of the left column (data column) shown in FIG. 12 is [5,1,1,1,1,1,5,2] (the sum of these is 17, that is, the 17 modules described above) The line width array as shown in FIG.

Next, the determination means 15c of the T sequence measurement unit 15 determines whether or not the cluster number calculated from the line width normalized T sequence matches the theoretical value.
(Code word conversion in low-level decoding)

  Finally, the codeword conversion unit 16 of the low level decoding unit 13d1 performs codeword conversion (step S133). More specifically, the code word corresponding to the T sequence obtained in step S132 is obtained by reading from the table memory 16a storing the code word conversion table and referring to the table memory 16a. That is, 929 codes in which low-level decoding processing is performed and code words each having a line width of four bars and a line width of four spaces constituting one column are numbered from 0 to 928. Once converted to the intermediate information indicated by.

Here, in the first embodiment, in the X sequence measurement process (step S131), an optimum one can be selected from a plurality of X sequences obtained from the threshold method and the second derivative method.
(About the threshold method)

FIG. 14 is a diagram in which a part of a scanning reflectance waveform indicating the luminance value of each pixel obtained by scanning a line waveform obtained by averaging a plurality of lines L included in a row in a certain row is extracted. This corresponds to a row and a column with a stack barcode 21. Whether a bar or a space is determined based on whether the luminance value of each pixel is higher or lower than a first (fixed) threshold AV, specifically, between a start point XS and an end point XE of a certain column, The X sequence is valid if the total number of spaces is eight. In the example of FIG. 14, the X sequence (bar line width and space line width) is measured based on the first (fixed) threshold AV, and the X sequence [12, 6, 3, 18, 3, 4, 12 is measured. , 11] was obtained. Here, the first (fixed) threshold value AV is an average value of all image data including the line widths of bars and spaces obtained by scanning an averaged line waveform in a certain column in a certain row. Specifically, in the row and the column shown in FIG. 14, the luminance values of all the measured pixels are added, and the value obtained by dividing the added value by the total number of pixels, that is, the average luminance value per pixel (fixed). ) The threshold value AV is set. Therefore, the first (fixed) threshold value AV is a constant value as shown in FIG. Thus, since the first (fixed) threshold value AV can be obtained by a single process, the processing time can be shortened.
(About nth-order differentiation method)

  FIG. 15 is a diagram for explaining the nth-order differentiation method. FIG. 15A is a diagram in which a part of a scanning reflection waveform indicating a luminance value of each pixel obtained by scanning a line waveform obtained by averaging a plurality of lines L included in a row in a certain row is extracted. It is. FIG. 15B is a waveform diagram showing a difference in luminance value obtained by first-order differentiation of the scanning reflection waveform shown in FIG. 15A. FIG. 15C is a diagram showing further differentiation (second-order differentiation) of the waveform in FIG. It is a wave form diagram which shows the difference of the brightness value which carried out.

The second-order differential method is to secondarily differentiate the scanning reflectance waveform indicating the luminance value of each pixel to obtain a second-order derivative value, and the zero-cross point of this second-order derivative value (that is, the inflection point of the original waveform). Is a boundary between a bar and a space. Compared with the method of detecting the peak of a waveform that is easily affected by the resolution, such as the threshold method, this secondary differentiation method obtains the boundary between the bar and the space by obtaining the waveform change point. Therefore, there is an advantage that the processing becomes simple. Note that the resolution of the scanning reflectance waveform shown in FIG. 15A is lower than the resolution of the scanning reflectance waveform shown in FIG.
(About X sequence measurement processing)

  FIG. 13 is a flowchart showing the flow of the X sequence measurement process (step S131). In this X sequence measurement process, as described above, in a certain row, a line waveform obtained by averaging a plurality of lines L included in the row is scanned, and the measured luminance value of each pixel is set to a predetermined threshold Th. If the value is higher, it is determined as a space, and if the luminance value of each pixel is lower than the threshold Th, it is determined as a bar by determining as a bar. Furthermore, based on such determination, the line width of the bar and the line width of the space are measured, and the line width is created. Since the data column in the stacked barcode 21 has four bars, four spaces, and the total number of bars and spaces is 8, it is formed between adjacent intersections of the scanning reflectance waveform and the fixed threshold AV. The process is repeated until the total number of line widths reaches 8, and the threshold value Th is varied.

  In the first embodiment, as shown in FIG. 13, first, the fixed threshold value calculation means 14a1 obtains an X sequence by the (fixed) threshold value method (steps S151 to S158).

  That is, when I = 1, the first fixed threshold value AV is set as the threshold value Th (step S151). The fixed threshold AV will be described later.

  Next, the X sequence when the first (fixed) threshold Th = AV is measured (step S152), and it is determined whether or not the total number of bars and spaces at that time is eight (step S153). If it is determined in step S153 that the total number is 8, the X sequence measurement is assumed to be successful, the X sequence at this time is validated (step S154), and this is stored as the X sequence 1 (step S158). On the other hand, if it is determined in step S153 that the total number is not eight, it is determined whether or not the process of the specified number of times I (I is an integer of 2 or more) has been repeated, for example, whether I = 3 has been reached (Step S155).

  If it is determined in step S155 that I = 3 is not satisfied, the specified number of times I is incremented by 1, and in the first embodiment, the second (fixed) obtained by subtracting AV / 4 from the first (fixed) threshold Th. A threshold value Th is set (step S156), and the processes of steps S152 and S153 are repeated. In FIG. 14, the symbol L <b> 1 indicates the nth (fixed) threshold Th obtained by subtracting AV / 4 from the (n−1) th (fixed) threshold Th. AV / 4 is a value determined from an empirical rule, and is not limited to this.

  Based on the newly obtained second (fixed) threshold Th, the X sequence is measured again (step S152). If the total number is determined to be 8, the X sequence is validated (step S154), and the total number is If it is not determined that the number is 8, the specified number of times I is not 3 (step S155), so that I is incremented by 1 and a new third (subtracting AV / 4 from the second (fixed) threshold Th is obtained. The (fixed) threshold value Th is set (step S156), and the X sequence is measured again based on the third (fixed) threshold value Th (step S152). In the first embodiment, if the total number is not determined to be 8 even after the process of the specified number of times I = 3 is repeated, the X sequence is invalidated (step S157) and stored as the X sequence 1 (step S158). ).

  When the X sequence measurement processing by the (fixed) threshold method is completed in this way, the n-th order differential calculation means 14a3 obtains the X sequence by the second order differential method (steps S159 to S166).

  That is, when I = 1, the zero cross point SL is set (step S159).

  Next, the X sequence at this time is measured (step S160), and it is determined whether or not the total number at that time is eight (step S161). If it is determined in step S161 that the total number is 8, the X sequence measurement is considered successful, the X sequence at this time is validated (step S162), and this is stored as the X sequence 2 (step S166). On the other hand, if it is determined in step S161 that the total number is not 8, it is determined whether or not the process of the specified number of times I (I is an integer of 2 or more) has been repeated, for example, whether or not I = 3 has been reached. (Step S163).

  If it is determined in step S163 that I = 3, 1 is incremented to the specified number of times I, 1 is incremented to the zero cross point SL (step S164), and the processes of steps S160 and S161 are repeated.

  Based on the newly obtained zero cross point SL, the X sequence is measured again (step S160). If the total number is determined to be 8, the X sequence is validated (step S162), and the total number is determined to be 8. If not, the specified number of times I is not 3 (step S163), and 1 is incremented to I and 1 is incremented to the zero-cross point (showing line) SL (step S164) second zero-cross (showing line) SL2 is calculated, and the X sequence is measured again based on the second zero-cross point (a line indicating) SL2 (step S160). In the first embodiment, if the total number is not determined to be 8 even if the processing of the specified number of times I = 3 is repeated, the X sequence is invalidated (step S165) and stored as the X sequence 2 (step S166). ). The increment number is 1, but is not limited to this.

  FIG. 16 is a flowchart showing the flow of the T sequence measurement process (step S132) in the first embodiment. In this T-sequence measurement process, the sequence conversion means 15a of the T-sequence measurement unit 15 performs X sequence 1 (X sequence obtained from the fixed threshold method) and X sequence 2 (X sequence obtained from the second derivative method). Are converted into T sequences, the line width normalizing means 15b of the T sequence measuring section 15 normalizes the line widths of these T sequences, and the determining means 15c of the T sequence measuring section 15 calculates from the respective T sequences. The obtained cluster number is compared with the theoretical value of the cluster. If they are equal, the T sequence is validated, and if they are not equal, the T sequence is invalidated.

  Here, the cluster number is prepared to minimize the probability of error occurrence in the stack barcode 21 shown in FIG. 3, and is present in all code words other than the start pattern and stop pattern. ing. Specifically, according to the PDF417 standard, the cluster number repeatedly appears every third row, the cluster number “0” in row 1, the cluster number “3” in row 2, and the cluster number “6” in row 3. Is used. In other words, the cluster numbers “0”, “3”, and “6” are the theoretical values of the cluster numbers.

  Next, specific numerical values will be described.

  First, the X sequence 2 obtained by the secondary differentiation method is taken out (step S171) and converted into a T sequence (step S172). This is obtained by adding two adjacent Xi and Xi + 1 in X = [Xi, Xi + 1, Xi + 2,...], And is an X sequence 2 obtained by the second-order differential method, for example, FIG. ), X = [12,5,5,16,4,4,12,10] is measured, and when this X is converted into a T sequence, T = [17, 10, 21, 20, 8, 16, 22].

  Next, a cluster number is calculated from the T sequence (step S173). In the calculation of clusters, first, the T sequence is normalized by TN (i) = T (i) * 17 / Wc (i = 0, 1,..., 7). In the above example, the line width array is as shown in TN = [4, 3, 5, 5, 2, 4]. Next, the cluster number K = (TN (0) −TN (1) + TN (4) −TN (5) +9) mod 9 (where “mod 9” means the remainder after dividing by 9). The cluster number K is calculated by the following formula. In the above example, K = (4−3 + 2−4 + 9) mod 9 = 8.

  Next, it is determined whether or not the cluster number calculated in step S173 is equal to the theoretical value of the cluster number corresponding to the row being processed (step S174). If the row being processed is the first row of the stack barcode 21, the theoretical value of the cluster is “0”. Accordingly, since the cluster number “8” does not match the theoretical value “0” of the cluster number, the T sequence is invalidated and the process proceeds to step S176. On the other hand, if the cluster number matches the theoretical value, the T sequence is validated (step S175).

  If the cluster number does not match the theoretical value in step S174, the X sequence 1 obtained by the fixed threshold method is extracted (step S176) and converted into a T sequence (step S177). X sequence 1 obtained by the (fixed) threshold method, for example, X = [12, 6, 3, 18, 3, 4, 12, 11] is measured from the scanning reflectance waveform shown in FIG. , T = [18, 9, 21, 21, 7, 16, 23].

  Next, a cluster number is calculated from the T sequence (step S178). In the above example, it is normalized as TN = [4, 2, 5, 5, 2, 4], and K = (4-2 + 2-4 + 9) mod 9 = 0.

  Next, it is determined whether or not the cluster calculated in step S173 is equal to the theoretical value of the cluster corresponding to the row being processed (step S179). In the above example, since the cluster number “0” matches the theoretical value “0” of the cluster number, the T sequence is validated (step S175). On the other hand, if the cluster number does not match the theoretical value, the T sequence is invalidated (step S180).

  As described above, in the first embodiment, two types of low-level decoding schemes, ie, the fixed threshold scheme and the second-order differential scheme, are provided, and by using them properly according to the level of the resolution of the imaging optical system, stacking is possible. Even if the scanning reflectance waveform of the mold barcode 21 fluctuates, an optimum X sequence (X sequence 1 or X sequence 2) can be obtained.

When the optimum X sequence is obtained, low-level decoding processing is performed, and codeword conversion processing (step S133) is performed. Specifically, a code word corresponding to the validated T sequence is acquired from a predetermined conversion table. That is, a low-level decoding process is performed, and a code word consisting of the line width of four bars and the line width of four spaces constituting one column is assigned 929 codes numbered from 0 to 928. Once converted to the intermediate information indicated by.
(High-level decoding)

When the codeword conversion is completed, general high-level decoding is performed (step S37), and the decoding process is completed. Further, the data for which the decoding process has been completed is output to the upper control apparatus together with the stack barcode information and the like. Specifically, the high-level decoding process is once converted into intermediate information indicated by 929 codes numbered from 0 to 928, and then decoded (decoded) into the final language based on a predetermined rule. It has become so. As the final language, PDF417 includes ASCII characters, binary expressions, and other characters.
(Main effects of the embodiment)

  As described above, in this embodiment, even when the scanning reflectance waveform of the stack barcode 21 fluctuates, the stack barcode 21 can be read with high accuracy. Specifically, when the resolution of the imaging optical system has reached a certain level, the threshold is determined by detecting the peak value of the scanning reflectance waveform of the stack barcode 21 (fixed). Low level decoding can be adopted. On the other hand, if the resolution of the imaging optical system does not reach a certain level, the scanning reflectance waveform of the stack barcode 21 will fluctuate, or it will be difficult to detect an accurate peak value. By taking the difference between the luminance values, it is possible to adopt low-level decoding by the second-order differential method that can extract the output change of the edge (boundary) portion between the bar and the space with high accuracy.

  As a result, it is difficult to be affected by fluctuations in the scanning reflectance waveform of the stack barcode 21, and the stack barcode 21 can be read with high accuracy.

In addition, even when the medium conveyance speed increases or decreases during imaging, there is an advantage that it is easy to deal with speed fluctuation by using both types ((fixed) threshold method and second-order differential method) of low-level decoding.
[Other Embodiments of the Invention]

  In the first embodiment described above, at the time of low-level decoding in the stack type barcode reading method, as shown in FIG. 16, low-level decoding based on X-sequence measurement processing by the second-order differential method (steps S171 to S174). In the above description, the low-level decoding (steps S176 to S179) based on the (fixed) threshold method X sequence measurement process is performed. However, if necessary, the order is changed, and low-level decoding based on the (fixed) threshold method X-sequence measurement processing is performed, and then low-level decoding is performed based on the second-order differential method X-sequence measurement processing. You may make it do.

  In the first embodiment described above, when executing the X-sequence measurement process of the stack barcode reading method, the fixed threshold method for obtaining the average of the luminance values of the pixels constituting the scanning reflection waveform is used as the threshold method. Although the case where it is adopted has been described, a threshold method other than the fixed threshold method (for example, a local threshold method) can be used instead or in combination. When the local threshold method is adopted, the local threshold value calculation unit 14a2 of the X sequence measurement unit 14 calculates the local threshold value. As shown in FIG. 17, this local threshold method is a method in which a local threshold based on the average value of the maximum peak value and the minimum peak value of bars and spaces adjacent to each other in image data is used as the boundary between the bar and the space. . This local threshold method can detect the line width between the bar and the space more accurately than the fixed threshold method even when the fluctuation of the peak value occurs due to uneven brightness or dirt of the corrected image data. There are advantages.

  In the first embodiment described above, the case where the second-order differentiation method (that is, n = 2) is adopted as the n-th order differentiation method when executing the X-sequence measurement process of the stack barcode reading method is described. However, instead of the secondary differential system, the primary differential system (that is, n = 1) can be substituted or used in combination.

  In the first embodiment described above, the stack bar code tilt correction (step S33) of the stack bar code reading method converts the image data of the stack bar code 21 into corrected image data with a tilt angle of zero. explained. However, when it is guaranteed that the image data of the stack barcode 21 is not tilted, or when the image data of the stack barcode 21 is tilted, this stack can be dealt with by another angle correction means. It is also possible to omit the type bar code tilt correction.

  In the first embodiment described above, the stack type barcode 21, that is, a kind of two-dimensional barcode, has been described. However, the present invention can be similarly applied to a one-dimensional barcode.

  Furthermore, in Embodiment 1 described above, the present invention is applied to a combination with a one-dimensional imaging device and a linear transport mechanism. However, the present invention is not limited to this, and for example, a two-dimensional CCD, The present invention may be applied to a combination of an area sensor such as a CMOS imager and a subject support mechanism.

  The present invention can be suitably applied to a stack type barcode reader in which the scanning reflectance waveform of the stack type barcode may fluctuate due to the resolution of the imaging optical system.

DESCRIPTION OF SYMBOLS 1 ... Stack type barcode reader 2 ... Record carrier 11 ... Imaging device (imaging part)
11a: Image scanner 11b: Card transport mechanism 12: Image memory 13: Data processing device (data processing unit)
13a: Position detection unit 13b: Inclination correction unit 13c: Structural analysis unit 13d: Decoding unit 13d1: Low level decoding unit 13d2: High level decoding unit 14: X sequence measurement unit 14a: Threshold calculation means 14a1... Fixed threshold calculation means 14a2... Local threshold calculation means 14a3... Nth-order differential calculation means 14b... Line width measurement means 15 ... T sequence measurement unit (T sequence measurement means)
15a: Sequence conversion means 15b: Line width normalization means 15c: Determination means 16: Code word conversion unit (code word acquisition means)
16a …… Table memory 21 …… Stacked barcode

Claims (5)

An imaging unit that captures a stack barcode in which character information is encoded into a barcode and displayed in a bar and a space;
An image memory for storing the image data of the stack barcode captured by the imaging unit;
A data processing unit having a decoding unit that decodes the barcode of the stack barcode into character information based on the image data;
The decoding unit
Threshold calculating means for calculating a threshold for measuring the line width of the bar and space of the stack barcode according to the luminance values of a plurality of pixels constituting the image data;
N-order differential calculation means for differentiating a scanning reflectance waveform indicating luminance values of a plurality of pixels constituting the image data to calculate an n-order differential value (n is 1 or 2);
Line width measuring means for measuring the line width of the bar and space of the stack barcode based on the threshold value or the nth-order differential value;
T-sequence measurement means for measuring a T-sequence from the line width of the bar and space of the stacked barcode measured by the line width measurement means;
A stack type barcode reader having code word acquisition means for acquiring a code word corresponding to the T sequence measured by the T sequence measurement means ,
The T sequence measuring means determines the validity of the T sequence measured based on one of the threshold and the n-th order differential value, and when the T sequence is not valid, the threshold and the n-th order are measured. A stack type barcode reader characterized by determining the validity of a T-sequence measured based on the other of the differential values .   2. The stacked bar code reader according to claim 1, wherein the n-th order differential value is a second-order differential value, and a zero cross point of the second-order differential value is used as a boundary between the bar and the space.   The threshold is one of a fixed threshold based on an average value of the image data and a local threshold based on an average value of the maximum peak value and the minimum peak value of the bar and the space adjacent to each other in the image data. The stacked bar code reader according to claim 1, wherein the stack bar code reader is provided. When decoding the barcode of the stack barcode into the character information based on the image data obtained by imaging the stack barcode displayed in the bar and space in which the character information is encoded into the barcode,
A threshold value calculating step for calculating a threshold value for measuring the line width of the bar and space of the stack barcode according to the luminance values of a plurality of pixels constituting the image data;
An n-th order differential calculation step of differentiating a scanning reflectance waveform indicating luminance values of a plurality of pixels constituting the image data to calculate an n-th order differential value (n is 1 or 2);
A line width measuring step of measuring the line width of the bar and space of the stack barcode based on the threshold value or the nth-order differential value;
A T-sequence measurement step of measuring a T-sequence from the line width of the bar and space of the stacked barcode measured in the line width measurement step;
A code word acquisition step of acquiring a code word corresponding to the T sequence measured in the T sequence measurement step ,
In the T sequence measurement step, the effectiveness of the T sequence measured based on one of the threshold value and the n-th order differential value is determined. If the T sequence is not valid, the threshold value and the n-th order value are determined. A stack type barcode reading method, wherein the validity of a T sequence measured based on the other of the differential values is determined .   5. The stack type bar according to claim 4, wherein the validity of the T-sequence is determined based on whether or not the cluster number calculated from the T-sequence measured in the T-sequence measurement step is equal to a theoretical value. Code reading method.