JP2722434B2 - Optical character reader - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to an optical reading device, in particular, reading a character by holding a scanner with a hand on a line on which a character such as a price tag is written, The present invention relates to a hand-held OCR (optical character reader). [Prior Art] A POS (Point Of Sales) system that collects sales information for each item in a supermarket or a department store and manages inventory has become widespread. As an OCR that can be used in this POS system, there is a hand-held optical reading device that can read characters in the field of view simply by applying a scanner to paper to be read (for example, Japanese Patent Application No. 60-79).
082). FIG. 2 shows a typical configuration of a hand-held optical character reader. Reference numeral 1 denotes a scanner which is applied to the paper 3 by hand 2 and reads letters, symbols, etc. written on the paper 3 by pressing a push button switch 19. The sheet 3 is, for example, a price tag on which information is described in a POS system. 4 is a light source, 5 is a lens system, and 6 is an image sensor. A driving and binarizing circuit 7 converts an analog signal which is an output signal of the image sensor 6 into a binarized signal, and sends it to a screen memory 8 in the main body A of the character reading apparatus. The screen memory 8 stores binarized data of almost the entire field of view of the image sensor 6. FIG. 3A illustrates the binarized data of the image sensor 6. The horizontal (X) x vertical (Y) size is p x
It is an image sensor of q pixels, and captures characters in the field of view. Characters and symbols are identified by the character identification circuit 13. Since the character identification circuit 13 identifies one character at a time, it is necessary to extract data for one character from the screen memory 8.
The one-digit extraction circuit 9 extracts data corresponding to m × q pixels, which is the processing capability of the one-character extraction circuit 11, from the screen memory 8 and stores it in the one-digit memory 10. One character extraction circuit 11
Is the processing capacity of the character identification circuit 13 from the single digit memory, mx
Data corresponding to n pixels is extracted and stored in the one-character memory 12. In FIG. 3 (a), first, the single digit cutout circuit 9
= 1 to X = m and data from Y = 1 to Y = q are taken out of the screen memory 8 and transferred to the single digit memory 10 (FIG. 3 (b ₁ )). The single digit cutout circuit 9 looks at the contents of the single digit memory 10 and includes a range including a character image (in this example, Y = 11 to Y = 11 +
The n lines of n-1) is transferred to the character memory 12 (FIG. 3 (c _1)). When a character is stored in the one-character memory 12, the character is recognized by the character identification circuit 13. Next, data from X = 2 to X = m + 1 and data from Y = 1 to Y = q are taken out from the screen memory 8 and transferred to the single digit memory 10 (FIG. 3 (b ₂ )). Then, the image in the range including the character image is transferred to the one-character memory 12. Hereinafter, similarly, the positions to be taken out to the screen memory 8 are sequentially shifted and transferred to the one-digit memory 10, the image including the character image is transferred to the one-character memory 12, and the character identification circuit
Recognition for one line is performed by performing the processing in 13. FIG. 4 shows how to determine the range to be transferred from the one-digit memory 10 to the one-character memory 12. First, a horizontal OR is obtained for each row of the one-digit memory 10. The horizontal OR is an operation that pays attention to one row in the horizontal direction, sets 1 if there is a black pixel in that row, and sets 0 if there is no black pixel. If the black output of the sensor is expressed as 1 and the white output is expressed as 0, the result of the horizontal OR is nothing but the result of ORing the pixels of one row. Therefore, this operation is called horizontal OR. Then, as shown in FIG. 4 (b), the result of the horizontal OR is black in the portion where the character exists, as shown in FIG. The range of transfer from single digit memory to single character memory is, for example, Y = 13
If the horizontal OR becomes black, then from n = 11 to n pixels including white above the character. Through the above processing, characters and symbols included in the field of view of the sensor 6 can be read. The read result is sent to the host computer B (such as a POS register) for use. [Problems to be Solved by the Invention] In the conventional optical character reader, the identification result often includes a reading error or a reading error,
The identification result cannot be used reliably. In order to reduce reading errors and reading errors, there is a method in which a plurality of identification results are obtained for the same reading target and a majority decision is made. However, in the conventional optical character reading device,
It has been found that simply adding a majority decision process cannot sufficiently reduce read defects and read errors. FIGS. 5 and 6 show a problem in the case where a majority decision processing unit is simply added to the conventional optical character reading apparatus. FIG. 5 is a configuration diagram in a case where three identification results are obtained for the same reading target and a majority decision is taken. By pressing the switch 19, the image of the character written on the paper 3 captured by the image sensor 6 is sent to the screen memory 8, and the single digit cutout circuit 9, single digit memory 10, single character cutout circuit 11, single character memory 12 An identification result is obtained through each part of the character identification circuit 13. This process is repeated three times. Then, the identification result buffer # 114 has the first
The second identification result is stored in the identification result buffer unit # 215, and the third identification result is stored in the identification result buffer # 316. When the identification result is set in all of the buffers # 1 to 314 to 16, the majority decision processing section 17 takes a majority decision and sends the result to the host computer B. FIG. 6 shows a majority decision example. FIG. 6A shows a character line described on a sheet. (B ₁ ), (b ₂ ), and (b ₃ ) show the first, second, and third identification results, respectively.
“?” Is a symbol indicating reject, and indicates that the character could not be read (read failure). (B ₄ ) is a majority decision of (b ₁ ), (b ₂ ), and (b ₃ ), and the same character line as that of (a), that is, the correct answer is obtained. (C ₁ ),
(C ₂ ) and (c ₃ ) are other examples of the identification results.
The second and third identification results are shown. (C ₁ ) is correct, but (c ₂ ) is missing “1”,
In (c ₃ ), “2” is omitted. (C ₄ ) is (c ₁ ),
These are the majority decision results of (c ₂ ) and (c ₃ ), and no correct answer was obtained unlike (b ₄ ). That is, in the majority decision processing configuration shown in FIG. 5, it is difficult to obtain a correct answer when characters are missing. SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and aims to improve the reading performance of an optical character reading device by realizing a majority processing method capable of finding a correct answer even when characters are missing. I have. [Means for Solving the Problems] As shown in FIG.
A case is shown in which three identification results are obtained for the same reading target and a majority decision is made on them. By pressing the switch 19, the image of the character written on the paper 3 captured by the image sensor 6 is sent to the screen memory 8 in the main body A of the character reading device, and the single digit cutout circuit 9 ', the single digit memory 10, An identification result is obtained through each part of the one-character extraction circuit 11, the one-character memory 12, and the character identification circuit 13. The x-coordinate of the character detected by the one-digit extraction circuit 9 'is stored in the identification result buffer # 114' together with the identification result. Set. This process is performed three times for one line, and the x coordinate of the character and the identification result are stored in identification result buffers # 1 to # 314 'to 16', respectively. After the x-coordinate a of the character and the identification result are stored in all the buffers # 1 to # 314 'to 16', the x-coordinate a of each character stored in each of the buffers # 1 to # 314 'to 16' is converted to a digit. It is sent to the matching processing unit 18. The digit alignment processing unit 18 performs a process of selecting one or a plurality of identification results at the same digit based on the character positions (coordinates) corresponding to the respective identification results obtained by the above-described three identification processes. . The selected identification result for each digit is determined by taking a majority vote in the majority decision processing unit 17, and the determined identification result is sent to the host computer B. [Operation] FIG. 7 shows a majority decision example according to the present embodiment. (A) is a row of a character string, and three times identification processing is performed on this character string. (B) Three identification results (b ₁ ) in the figure,
Even if a digit is missing in (b ₂ ) and (b ₃ ) and a character is missing, a majority decision based on only the identification result for the same digit as shown in (b ₄ ) is performed by the digit matching process. Since the processing can be performed, a correct majority decision result (b ₄ ) can be obtained as compared with a majority decision result (c ₄ ) in which digit alignment processing is not performed as in the conventional example shown in FIG. Embodiment FIG. 1 shows a first embodiment of the present invention. Reference numeral 1 denotes a switch for reading characters written on the paper 3 by placing the switch on the paper 3 with the hand 2 and pressing the switch 19. 4 is a light source, 5 is a lens system, 6 is an image sensor, and 7 is a control binarization circuit. Reference numeral 8 denotes a screen memory provided in the main body A of the character reading device. A one-digit cut-out circuit 9 'cuts out a character string forming one digit from the binarized signal stored in the screen memory 8, and outputs each character. Is detected. The one-digit binary signal extracted by the one-digit extraction circuit 9 'is stored in the one-digit memory 10, and the identification result obtained through the one-character extraction circuit 11, the one-character memory 12, and the character identification result circuit 13 is extracted by one digit. It is stored in the identification result buffer # 114 'together with the x coordinate a of the character obtained by the circuit 9'. After finishing the identification result for one digit, the image of the character written on the paper 3 is taken in the image sensor 6 again, the identification result is performed, and the x coordinate a of the character and the identification result are stored in the identification result buffer # 215 '. . Identification result buffer # 31
Repeat for 6 '. X coordinate a of the stored character
The digitizing processing unit 18 performs digit matching processing of the identification result based on the result, and the majority of the digitized identification result is determined by the majority decision processing unit 17 to determine one identification result for one digit. The result is sent to the host computer B. FIG. 8 shows the second embodiment, and FIG. 9 is a schematic flowchart showing the operation of the microprocessor 24 in the second embodiment. X of the character sent from the single digit cutout circuit 9
The coordinates a and the identification result sent from the character identification circuit are received by the microprocessor 24 and then stored in the RAM 26.
The x-coordinate a and the identification result of each obtained character are temporarily stored in the RAM 26 with respect to the image data captured three times at staggered times, and then the microprocessor 2
4 and is stored in the ROM 25 to check whether or not the range of the x-coordinate a of the character recognized as having the same digit is satisfied. If one or more identification results are recognized as having the same digit, To be elected. If there is only one discrimination result recognized as having the same digit, the discrimination result is output as the discrimination result of the character at that digit, and if there are a plurality of discrimination results, a majority decision is taken and only the most discrimination result is output. The digit alignment process using a microprocessor is shown in more detail in FIG. Now, the identification result on the first screen has one character, and the i-th (i = 1, 2,... I) character and its x coordinate are denoted by ▲ c ⁽¹⁾ _i ▼ and ▲ x ⁽¹⁾ _i ▼, respectively. And the identification result on the second screen has a J character,
= 1,2, ... J) and its x-coordinate are ▲ c ⁽²⁾ _j
▼, ▲ x ⁽²⁾ _j ▼, the identification result on the third screen has K characters, and the k-th (k = 1, 2,... K) character and its x coordinate are ▲ c ⁽³⁾ _k ▼, ▲ x ⁽³⁾ _k ▼, and the characters of the identification result of the first screen, second screen, and third screen after digit alignment are ▲ y ⁽¹⁾ _l ▼, ▲ y ⁽²⁾ _l ▼, ▲ y ⁽³⁾ _l ▼ (l
= 1,2,... L). In the flowchart of FIG. 13, the initialization is performed in. Is a process in which the end of the identification result of each screen is previously indicated by a special large value M. M is set to a value larger than the sum of the maximum value of the x coordinate that can be obtained as a result of the identification processing and the value of W described later. Is the minimum of the coordinates ｘx ⁽¹⁾ _i ▼ of the i-th character, the coordinates ｘx ⁽²⁾ _j ▼ of the _j- th character, and the coordinates ｘx ⁽³⁾ _k ▼ of the k-th character This is processing for _{obtaining the} value x _min . If x _min is the same as M used in the processing, it is determined that the digit matching processing for all the identification results has been completed. W indicates the width of the x-coordinate that can be determined to be the same digit. Now, if ▲ x ⁽¹⁾ _i ▼ ≦ x _min + W holds, ▲
x ⁽¹⁾ _i ▼ is a process for entering the digit currently being processed,
That is, the process of (1) is performed. Is the ▲ c of the i-th character
^{(1) In this process,} _i ▼ is registered as the identification result ▲ y ⁽¹⁾ _l ▼ after digit alignment. Is ▲ c ⁽¹⁾ _i ▼
Is a process of increasing the value of. On the other hand, when ▲ x ⁽¹⁾ _i ▼ does not fall in the digit currently being processed, the processing is performed. Is a process for registering a symbol # indicating that no identification result is obtained in the corresponding digit in ▲ y ⁽¹⁾ _l ▼ (# is the same as the space between characters in FIG. 7B) Meaning). In the same manner as above, processing for the identification result of the second screen and processing for the identification result of the third screen are performed. Is a process for increasing the number l of characters of the identification result after digit alignment by one. Is a process of registering the number of characters (l-1) finally obtained as a variable L. In FIG. 9, after the "digit adjustment of three-screen identification result" is completed, "take majority decision of three-screen identification result"
Although a flowchart is shown, a process of performing digit alignment and majority decision for each digit is also possible. FIG. 10 is an example of processing when reading a plurality of lines at a time. If multiple lines enter the field of view of the image sensor,
The discrimination processing is performed three times from the one line located on the upper side, and the majority decision processing of the discriminated identification result is performed. After the majority decision processing for the upper one row is completed, the processing of the second and third rows is performed in order. FIG. 11 shows three outputs for one output from the image sensor.
This is an embodiment in which binarization is performed by two binarization circuits. The binarized signal corresponds to each of the binarized circuits,
Types are obtained and stored in the screen memories 8 ', 8 ", and 8.
The processing from the single digit cutout circuit 9 'to the character identification circuit 13 is as follows.
It is performed once for each of the screen memories 8 ', 8 ", 8 and the identification results obtained for each of the screen memories 8', 8", 8 are stored in the identification result buffers # 1 to # 31.
Stored in 4 'to 16'. Fig. 12 shows that the switch does not capture the image,
This is an embodiment in which a single digit cutout circuit 9 'detects the presence of a character string on a sheet to start the identification process. [Effect] According to the present invention, it is possible to realize an optical character reading device capable of correctly performing a majority decision even for missing characters. Since the reading performance can be improved by the majority decision processing, the operation and time required for rereading in the case of a reading error can be reduced, and the burden on the operator can be significantly reduced.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an optical character reading apparatus according to the present invention, FIG. 2 is an optical character reading apparatus according to the prior art, and FIGS. 4 (a) to 4 (c) are illustrations of a method for extracting one character, and FIG. 5 is a configuration in which a majority decision processing unit is added to an optical character reading apparatus according to the prior art. FIG. 6 (a)
FIGS. 7A to 7C are majority examples of the prior art, and FIGS.
(C) is a majority decision example according to the present invention, FIG. 8 is a configuration diagram when the present invention is implemented using a microprocessor,
FIG. 9 is a schematic flowchart, FIG. 10 is a majority decision example according to the present invention when a plurality of rows are read at one time, FIG.
FIG. 12 is another embodiment of the optical character reading apparatus according to the present invention, and FIG. 13 is a digit alignment processing flowchart. 1 ... Switch 2 ... Hand 3 ... Paper 4 ... Illumination light source 5 ... Lens 6 ... Image sensors 7, 7 '
7 ″, 7... Drive and binarization circuit, 8, 8 ′,
8 ", 8 ... screen memory, 9, 9 '... one digit cutout circuit, 10 ... one digit memory, 11 ... one character cutout circuit,
12: One character memory, 13: Character identification circuit, 14, 14 '...
... Identification result buffers # 1, 15, 15 '... Identification result buffers # 2, 16, 16' ... Identification result buffers # 3, 17 ... Majority decision processing unit, 18 ... Digit alignment processing unit, 19 ... Press Button switch, 24 ... Microprocessor, 25 ... ROM, 26
……RAM.

Continuation of front page    (72) Inventor Kazuo Fujiwaki               1-3-3 Shimaya, Konohana-ku, Osaka City Sumitomo               Inside Electric Manufacturing Co., Ltd. Osaka Works                (56) References JP-A-56-31176 (JP, A)                 Tokiko Sho 60-43556 (JP, B2)

Claims (1)

(57) [Claims] In a hand-held optical character reading device that optically reads characters, symbols, and the like, an image sensor configured by arranging photoelectric exchange elements in a plane and storing one or more lines of characters, symbols, and the like in a field of view, A binarization circuit that converts an analog signal output from the image sensor into a binarization signal corresponding to a character area and a background area, and a memory that stores the binarization signal output from the binarization circuit, Character position detecting means for detecting the position of each character, symbol, etc. from the binarized signal stored in the memory by examining the boundary between the background region and the character region in the binarized signal; From the signal, the identification processing means for identifying the characters, symbols, and the like, and a plurality of identification results obtained by performing character position detection and identification processing a plurality of times for one line or a plurality of lines of characters, symbols, and the like, Identification of each Digit matching processing means for selecting one or more identification results at the same digit position based on the character position corresponding to the result, and identifying results for each digit selected by the digit matching processing means. An optical character reading apparatus comprising: majority processing means for, in one case, outputting the identification result as the identification result of the character at that digit;