JP2514663B2 - Optical character reader - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention is designed to scan characters on a manuscript with a hand-held scanner (hereinafter, only characters will be described as a representative, but the same applies to symbols). A) optical character reader.

[Conventional technology]

POS (Point of Sales) systems, which collect sales information for individual items and manage inventory, have become widespread in supermarkets and department stores. Handheld optical character readers are often used in this POS system.

As such a device, the present applicant has applied for a patent for Japanese Patent Application No. 62-11083 and Japanese Patent Application No. 62-56293. A typical structure of a handheld optical character reader is shown in FIG.

In FIG. 2, reference numeral 1 denotes a scanner, which is used to read characters stored in a manuscript simply by applying it to the manuscript 3. The manuscript 3 is, for example, a price tag sheet used in the POS system. Reference numeral 4 is a light source, 5 is a lens system, and 6 is an image sensor, which requires a field of view of at least one line written on the paper 3, and in FIG. I'm doubling it. Reference numeral 7 denotes a control / binarization circuit which converts an analog signal, which is an output signal of the image sensor 6, into a corresponding binarized signal for each of a character area and a background area, and sends it to the screen memory 8.

9 to 13 are means for recognizing each character in the screen memory 8 and obtaining the position (x coordinate) of the character in the visual field.

The screen memory 8 stores binarized data of almost the entire visual field of the image sensor 6. FIG. 3A shows an explanation of the binarized data of the image sensor 6. An image sensor having a size of horizontal (X) × vertical (Y) of p × q pixels,
The characters in the field of view are captured.

Characters and symbols are recognized by the character identifying circuit 13, but since the character identifying circuit 13 recognizes character by character, it is necessary to retrieve data for one character from the screen memory 8.
The one-digit cutout circuit 9 takes out from the screen memory 8 data corresponding to m × q pixels, which is the processing capacity of the one-character cutout circuit 11, and stores it in the one-digit memory 10. Single character extraction circuit 11
Is the processing capacity of the character identification circuit 13 from the one-digit memory m ×
Data corresponding to n pixels is taken out and stored in the one-character memory 12.

In FIG. 3 (a), first, the single digit cutout circuit 9 is X.
The data from = 1 to X = m and from Y = 1 to Y = q are taken out from the screen memory 8 and transferred to the one-digit memory 10 (FIG. 3 (b ₁ )). The one-digit cutting circuit 9 looks at the contents of the one-digit memory 10 and includes a character image in the range (Y = 11 to Y = 11 + in this example).
n-1) n lines are transferred to the one-character memory 12. (Third
Figure (c ₁ )). When a character is stored in the one-character memory 12, the character is recognized by the character identification circuit 13. Next, the data from X = 2 to X = m + 1 and Y = 1 to Y = q are taken out from the screen memory 8 and transferred to the one-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. Thereafter, similarly, the positions taken out from 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
By performing the process in 13, the character of one line is recognized.

FIG. 4 shows how to obtain the range to be transferred from the single digit memory 10 to the single character memory 12. First, the horizontal OR is calculated for each row of the one-digit memory 10.

The horizontal OR is an operation in which one row in the horizontal direction is focused and 1 is set if there is a black pixel in the row, and 0 is set if there is no black pixel. Now, if the black output of the sensor is expressed as 1 and the white output is expressed as 0, the result of the lateral OR is nothing but the result of the logical sum of the pixels in one row. Therefore, this operation is called horizontal OR. Then, as shown in FIG. 4 (b), the result of the horizontal OR in that part of the character is black as shown in FIG. The range for transferring from one-line memory to one-character memory is, for example, Y = 13.
Therefore, if the horizontal OR becomes black, Y = 11 to n pixels including white above the character.

Through the above processing, the characters and symbols included in the field of view of the sensor 6 can be read. The process of scanning the image sensor 6 to store the image of the paper 3 in the screen memory 8 and recognizing each character in the screen memory 8 is performed three times,
Identification result buffers of 14, 15 and 16 showing the character recognition result and the position of the character in the visual field (x-coordinate when the one-digit cutting circuit 9 cuts the screen into the one-digit memory 10 when the character is recognized). It is stored in # 1, # 2, and # 3.

In the character recognition device of FIG. 2, the recognition rate is improved by taking a majority decision of the result of repeatedly recognizing one character. 14 to 18 are means for obtaining a majority of the recognition results of characters when the screen is captured and recognized three times. First, the x-coordinate value of the character and the recognition result stored in the recognition result buffers # 1 to # 3 are sent to the digit alignment processing unit 17. The digit alignment processing unit 17 associates character recognition results that can be determined to be the same digit based on the value of the x coordinate of the character.
The majority decision processing unit 18 obtains a final recognition result for the digit by taking a majority decision between the associated character recognition results.
An example of majority decision is shown in FIG. (A) is a line described in the manuscript. (B ₁ ) (b ₂ ) (b ₃ ) represent the recognition results in the first, second, and third recognitions, respectively. (B ₁₎ the lack of character of "1", (b ₂₎ In the "2"
The character “3” is missing in (b ₃ ) and the character “3” is missing in (b ₃ ). _The correct answer is obtained as in ₄ ). If a majority decision is made by simply associating from the beginning of the recognized character without using the x coordinate of the character, the correct answer cannot be obtained as in (c).

Reference numerals 19 to 23 in FIG. 2 are means for outputting the recognition result which is correctly read for one line on the original document only once. As long as the line is in the field of view, the letters are repeated and recognized, and the majority decision result is repeatedly output from 18. FIG. 6 shows the movement of the visual field and the change in the majority result when the scanner 1 is moved from top to bottom while being applied to the line “C1234567890”. At the position of (a ₁ ), all recognition results (b ₁ ) are rejected (indicated by unrecognizable :?). (A ₂₎ the position of the recognition result of (b ₂₎ also are all rejected as well. In (a ₃₎ position, it has been recognized entered in the field of view only the characters of "0". The majority result from (b ₁ ) to (b ₃ ) is like (c ₁ ). Here, with the character "0" is in preference to reject the character recognition result of the _{(b 3) (b 1)} (b 2). That is, the character recognition result is weighted rather than rejected. Further moving the field of view scanner is as _{(a 4) (a 5)} (a 6) the recognition result at the position of _{(b 4) (b 5)} (b 6) the majority result of (c _2). Similarly, as _{(a 7) (a 8)} (a 9) the recognition result at the position of _{(b 7) (b 8)} (b 9) majority result of (c _3).

From (18), (c ₁ ) (c ₂ ) (c ₃ ) is sequentially output.
Is a format check unit, and a majority format result obtained from 18 has a predetermined format (for example, C
The line beginning with must satisfy C) followed by 10 digits). Timer 20
Measure the elapsed time since the majority result was obtained from 18. If the majority result R _i satisfying the predetermined format is obtained, the previous register 21, the comparator 22, the output control unit 23,
It works as follows. First, in the comparator 22, R _i is compared with the content R _i−1 previously stored in the register 21. If the contents of R _i and R _i-1 do not match, the comparator 22
The signal W is output and the output control unit 23 determines R _i to be the recognition result R
Output as _LINE . If the contents of R _i and R _i-1 match, the NEW signal is not output from the comparator 22 and the output control unit 23 does not output R _i (discard it). On the other hand, the previous register 21 is R _i-1
After sending to the comparator 22, R _i is stored. Timer 20
The time after the majority result is obtained from 18 is measured, and the content of the previous register 21 is erased after the elapse of a predetermined fixed time T _CLR . The state of the previous register immediately after the power is turned on is the erased state. T _CLR is shorter than the time required to change the price tag (for example, 1 second), and is set to about 0.6 seconds, for example. The repetition cycle when the row is in the field of view and the majority result is repeatedly obtained from 18 is, for example, about 0.2 seconds.

The operations 19 to 23 when reading the price tag will be described with reference to FIGS. 6 and 7. (C ₁ ) Majority result "????????
Since "? 0" does not satisfy the predetermined format, nothing is output from the format check unit 9. Move the scanner from top to bottom (c ₂ ) majority result “C12345
67890 ”is obtained, the format check unit 19 outputs the majority decision result R _i because it satisfies a predetermined format. The comparator 22 compares R _i with the contents of the previous register R _i-1, but the contents of the previous register are erased immediately after the power is turned on, so the contents of R _i and R _i-1 always match. do not do. Therefore, the NEW signal is output from the comparator 22 to the output control unit 23, and "C1234567890" is output from the output control unit 23.
Is output as the line recognition result R _LINE . On the other hand, “C1234567890” is stored in the previous register 21. Next, the majority result of (c ₃ ) "C1234567890" is obtained, but (c
_{As in the case of 2} ), it is sent to the comparator 22 through the format check unit. However, since the content R _i-1 of the register 21 last time is “C1234567890” and matches R _i , the NEW signal is no longer output from the comparator, and the output control unit 23 outputs R
Not output as _LINE . As described above, the line recognition result R _LINE is output only once for each line in the field of view.

The operation of moving the scanner to read a plurality of lines will be described with reference to FIG. It is assumed that the format check unit 19 has registered the format of each line starting with C, N, and \. First, when the scanner is applied to the line of "C1234567890" with the price tag of (a), as described above, the line recognition result R _LINE is output only once. Next, when the scanner is moved down and the scanner is applied to the line of "N1234567890", when the line of "N1234567890" enters the field of view and the recognition result of "N1234567890" is obtained from the majority processing unit 18 for the first time. Is the previous register contents R _i-1 is “C123456789
Since it is "0", the NEW signal is output from the comparator 22, and "N1234567890" is output as the line recognition result R _LINE . After that, even if "N1234567890" is repeatedly output from the majority processing unit 18, it is not output as R _LINE because it matches the previous register contents. That is, "N123456
7890 ”is output only once. Similarly, "¥ 123,45
When the line "6." enters the field of view, "\ 123,456." Is output as the line recognition result R _LINE . In the price tag for POS according to JIS B9551, each line in one price tag has different contents. Therefore, as a result of the recognition time and the result R _i-1 recognized last time comparator 22 to compare the R _i, it can be determined whether or not reading the same line.

There are no characters in the field of view while the price tag is changed from (a) to (b). At this time, no output is obtained from the majority decision processing unit 18. Timer 20 is majority processing unit 18
Is measuring the elapsed time after outputting the recognition result,
If the state in which no output is obtained from each character recognition means for T _CLR or more continues because the price tag is changed, the contents of the register 21 are erased last time. Therefore, if you change the price tag to (b) and apply the scanner to the line "¥ 123,456."
6. ”is output. That is, even if the lines have the same content, different price tags can be continuously read. As can be seen from the above description, the timer 20 has a function of detecting the exchange of the manuscript (price tag).

[Problems to be solved by the invention]

Tilt scanner 1 to the row (with skew)
FIG. 8 shows the state when the line is applied. The skew angle θ is small, when the entire line as (a ₂₎ is a timing falling within the field of _{view, (a 1) (a 2} ) (a 3) of the recognition result _{(b 1) (b 2)} ( In the majority result (c) of b ₃ ), the recognition result of the line is correctly obtained.

FIG. 9 shows how the skew angle θ is increased. It is assumed that the skew angle θ is not so large that one character completely entering the visual field cannot be recognized. Although there is no timing for the entire row to enter the field of view due to the increased skew angle, the majority of the results (c ₁ ) appear when the right edge of the row is in the field of view and (a ₃ ) is the left edge of the line is in the field of view. The line recognition result is correctly obtained in).

Show the movement of the visual field and the recognition result when slowly moving the scanner from top to bottom with the skew angle θ increased.
Shown in Figure 10. Due to the large skew angle θ, there is no time for the entire row to enter the field of view. Also, since the scanner is moving slowly, it is possible to obtain the majority decision result of any adjacent recognition results (b _i ) (b _{i + 1} ) (b _{i + 2} ) (i is an integer from 1 to 8). The whole recognition result cannot be obtained correctly.

As can be seen from the explanation of FIG. 10, in the conventional optical character reading device, when the line is captured by slowly moving the scanner with a large skew angle so that the entire line does not fit in the field of view at one time, the entire line is read. Can not get the recognition result of. The present invention has been devised to solve this drawback, and it is an object of the present invention to enable the entire line of an optical character reader to be recognized even at a skew angle such that the entire line of view does not fit at once.

[Structure of Invention]

FIG. 1 shows a configuration example of an optical character reading device using the present invention. In the figure, the parts denoted by the reference numerals 1 to 23 have the same functions and configurations as the parts denoted by the same reference numerals in the conventional optical character reader of FIG. However, the digit alignment processing unit 17 'outputs the value of the x coordinate of the character after digit alignment in addition to the function of the digit alignment processing unit 17 in FIG. For the x-coordinate of the character after digit alignment, for example, the average value of the x-coordinate values of the characters that can be regarded as the same digit is used. Characters 8 to 18 are character recognition means for recognizing each character on the screen captured by the image sensor and outputting the position. The recognition result of each character in the field of view obtained from the character recognition means and its position information will be referred to as a recognition line in the following description.

Reference numerals 31, 32, and 33 are row detection means for detecting whether or not a part of the row exists in each area when a plurality of areas are set in the field of view. Although FIG. 1 exemplifies a case in which there are three row detecting means, if there are a plurality of row detecting means, a configuration with a number other than 3 is also possible. Reference numerals 34, 35 and 36 are row storage means for storing the recognized row when the corresponding row existence detecting means detects the row. 37 is a digit alignment processing means for associating characters with each other in the recognition line stored in each of the line storage means based on the position information. Reference numeral 39 is a line comprehensive processing means for obtaining one line by synthesizing the recognized lines to which the respective characters are associated by the digit alignment processing means. In the following description, the line obtained by the line total processing means will be referred to as a total line. Reference numeral 39 is a line interchange detecting means, and is a means for detecting that the lines captured by the image sensor 6 have been interchanged.

[Action]

When the scanner 1 is applied to a new row, the contents of the row storage means 34, 35 and 36 are erased by the function of the row interchange detection means 39. Below, the operations of the line detection means, the row storage means, the digit alignment processing means, and the line integration processing means will be explained using the example of FIG.

Now, as shown in (a), the field of view 30 is ₃ of S ₁ , S ₂ , and S ₃ .
It is divided into two areas, and the row detecting means # 1, # 2, and # 3 respectively determine whether or not there is a row in each area. When the line shown in FIG. 11 (b) is read while tilting the scanner, the positional relationship between the field of view and the line and the recognized line are (c ₁ ) (c ₂ ) (c ₃ ) and (d ₁ ) (d ₂ ), respectively.
It changes sequentially like (d ₃ ). The stored contents of the row storage means and the row total processing result when the recognized rows of (d ₁ ) (d ₂ ) (d ₃ ) are sequentially obtained are shown in (e).

Before the scanner is applied to the line, the line replacement detection means
39, the contents of the row storage means # 1, # 2, # 3 are in the erased state. First, the state when a line enters the area S ₃ of the visual field as in (c ₁ ) and the recognition line in (d ₁ ) is obtained is shown in the second stage. Row detection means that a row has entered the area of S ₃ #
3 is detected and the recognized line is stored in the line storage means # 3. The contents 101, 102, 103 stored in each row storage means are subjected to digit alignment processing by the digit alignment processing means 37 (10
Four). In the first stage, since 101 and 102 are in the erased state, the total row 105 obtained by the row total processing is the same as 103.
Then, the characters of the comprehensive line obtained at 105 are sent to the format check unit 19. Now, it is assumed that the format check unit is registered as corresponding to FIG. 11B, that the character "V" is followed by twelve numbers. Since 105 does not match the predetermined format, nothing is output from the format check unit.

Next, as shown in FIG. 11 (c ₂ ), the state when the S ₂ region of the visual field is entered and the recognition line of (d ₂ ) is obtained is shown in the first stage. The row detection means # 2 detects that the row has entered the area S ₂ , and the recognized row is stored in the row storage means # 2. At this time, the contents of the row storage means # 1 and # 3 remain the same as in the first stage. Contents 106, 1 stored in each row storage means
Digit alignment processing means 37 performs digit alignment processing for 07 and 108 (109). In the line total processing, the character recognized in each digit is selected to obtain the total line 110. Then, the characters in the 110 total lines are the format check section 19
Sent to Since 110 does not match the predetermined format, nothing is output from the format check unit.

Next, as shown in FIG. 11 (c ₃ ), a line enters the area S ₁ of the visual field, and the state when the recognition line (d ₃ ) is obtained is shown in the first stage. The row detecting means # 1 detects that the row has entered the area of S ₁ , and the recognized row is stored in the row storing means # 1. First,
In the same manner as in the step, the digit alignment processing means 37 performs digit alignment processing on the reference numerals 111, 112, and 113 (11
Four). In the line total processing, the character recognized in each digit is selected to obtain the total line 115. The entire line is obtained in the stage 1 general line, at this time the format check section 1
9, output through the output control unit 23 (Fig. 1, R _LINE ).

By the above processing, when the scanner is applied to a line on the document and moved, the recognition result of the entire line can be obtained even if the entire line does not enter the visual field at once. When the scanner is further moved to read the next line, the line exchange detecting means 39 detects the line exchange and the line storing means #
The contents of 1, # 2 and # 3 are erased.

The operation when the entire row enters the field of view at once and the recognition row (g) is obtained as in (f) is as shown in (h). In each of the areas S ₁ , S ₂ , and S ₃ , the row detection means # 1, # 2, and # 3 detect that a row has entered the field of view.
The recognized line (g) is stored in each of the line storage means # 1, # 2, and # 3 (116, 117, 118). Then, the whole line is obtained as the result 120 of the digit alignment process / line integrated process.

〔Example〕

FIG. 12 shows an embodiment of the line-swapping means # 1, # 2, and # 3 and the row-swapping detection means. (A) shows the field of view 30 of the image sensor, which is divided into regions S ₁ , S ₂ , and S ₃ . Then, in each area, the range for detecting rows is set as exemplified by R ₁ , R ₂ , and R ₃ . The row detecting means detects that there is a row in this range. (B) is a circuit that implements the row detecting means and the row interchange detecting means. 12
Reference numerals ₁ , 122 and 123 are circuits for performing a horizontal OR operation in the range of R ₁ , R ₂ and R ₃ , respectively. Reference numerals 124, 125, and 126 are black length determination units that determine whether or not the length of consecutive black pixels in the horizontal OR result is within a predetermined range. Character lines are in the range R _i (i = 1,
2), the horizontal OR result indicates that the black pixels corresponding to the character are continuous for the height of the character, so the black length determination unit determines whether the character is in that area. Judged,
The result is output to the EXIST _i signal line. The EXIST _i signal is sent to the corresponding row storage means, and each row storage means gives the timing for storing the recognition row. 127 is EXIST ₁ , EXIS
This is an OR gate that takes the logical sum of T ₂ and EXIST ₃ . The output EXIST at 127 is true if there is a character in any of R ₁ , R ₂ and R ₃ . To detect the interchange of rows, it is sufficient to detect that the rows come out of the visual field, that is, EXIST becomes false. Reference numeral 128 is an inverter gate which inverts the logic of EXIST and creates a row interchange signal ALLCLR. ALLCLR
Signal is sent to all the row storage means to give a timing to erase the contents of the row storage means.

If the line is not in the field of view, the majority processing unit 18
Can not get the recognition line. Therefore, an embodiment is possible in which it is determined that the recognition line is not obtained and the line is replaced. FIG. 13 shows an embodiment in which the time when no recognition line is obtained is measured, and when it exceeds a predetermined time (T _L ), it is judged that the line is replaced. In Figure 13, the timer
130 outputs a signal of T _{LUP when} the time when no recognition line is obtained continues for T _L or more. _TL is set to be slightly larger (for example, 0.25 seconds) than the cycle (for example, 0.2 seconds) at which the repeated recognition row is obtained when the row is in the visual field.

In FIG. 13, the parts denoted by the reference numerals 1 to 23 have the same functions and configurations as the parts having the same reference numerals in FIG. Reference numeral 6 denotes an image sensor, which requires a field of view of at least one line of the characters written on the paper 3 as in the case of FIG. 2. In FIG. 13, the horizontal line is one line, and the vertical line is about three times as large as one character. I am trying. Also, if multiple lines are written on the original, complicated processing is required if multiple lines are in the field of view at one time. Therefore, to simplify the processing system, set the height of the image sensor field of view to the original. It is better to make it smaller than the line spacing of.

In FIG. 13, the line detection means # 1 to # 3, the row storage means, the digit alignment processing means, and the line total processing means are implemented using the microprocessor 131, the ROM 132, and the RAM 133. ROM
A program of the microprocessor 131 for implementing the line detection means, the digit alignment processing means, and the line integrated processing means is stored in 132. The row storage means 34 is implemented in the variable area set on the RAM 133.

FIG. 14 shows a schematic flowchart of the processing of the microprocessor 131. When the power of the optical character reader is turned on, the process starts from. Is a process for deleting the contents of the row storage area provided in the RAM. Timer 130 is T _LUP
In the process of determining whether or not a signal is being output, when the T _LUP signal is output, that is, when there is a row exchange, the process proceeds to. Is a determination as to whether a recognition line has been obtained. Is the process of reading the recognition line. Processing for detecting a line in each area when the field of view is divided into a plurality of areas and storing the recognized row in the row storage area corresponding to each area. Is. The flowchart of FIG. 14 shows the processing steps when the field of view is divided into three regions S ₁ , S ₂ , and S ₃ (FIG. 15 (a)). The process of storing the recognized row in the row storage area # 1 corresponding to the area of S ₁ is as follows. Whether the number of characters recognized (not rejected) in the area S ₁ of the recognition line is larger than the number of characters recognized in the area S ₁ of the contents stored in the line storage area # 1. This is a process for determining whether or not. , Indicate that the same processing as in the areas of S ₂ and S ₃ is performed, respectively. Is row storage # 1
Is a digit alignment process for finding the corresponding characters based on the position information of the recognized characters stored in contents # 3 to # 3. Is a process of performing an overall process based on the result of digit alignment performed in step S3 to obtain an overall line. Is a process for sending the total line to the format check unit 19.

16 to 18 show detailed flowcharts of the processing in FIG. In the following description, it is assumed that variables and constants are used as shown in FIG. That is, as shown in FIG. 15 (a),
The range of the x coordinate of each area of S ₁ , S ₂ , and S ₃ is from X ₁ to X ₂ ,
X ₂ to X ₃ , X ₃ to X ₄ . Also, as shown in the table of (b), the recognition line is N characters, and the character and position are
Represented by a ⁽⁰⁾ _n ▼, ▲ x ⁽⁰⁾ _n ▼, the content stored in the line storage area # 1 has an I character, and the character and position are ▲ a ⁽¹⁾ _i ▼,
It is expressed as ▲ x ⁽¹⁾ _i ▼ (i = 1, 2, ..., I), and S ₁
The number of recognized (non-rejected) characters in the area of is denoted by P ₁ . Similarly for row storage areas # 2 and # 3,
J, ▲ a ⁽²⁾ _j ▼, ▲ x ⁽²⁾ _j ▼, P ₂ and K, ▲ a ⁽³⁾ _K ▼, ▲ x
⁽³⁾ Use the symbols _K ▼ and P ₃ . The number of digits after digit alignment is represented by L, the position of each digit is y _l , and the combination of characters is ▲ b.
⁽¹⁾ _l ▼, ▲ b ⁽²⁾ _l ▼, ▲ b ⁽³⁾ _l ▼. The total number of characters is M and the characters are c _m .

FIG. 16 shows a detailed flowchart of the processing of FIG. Is a process for initializing a variable p that counts the number of recognized characters in the recognized line that enters the area of S ₁ . Is an iterative process for repeating the process of. Is
This is a process for determining whether or not ▲ a ⁽⁰⁾ _n ▼ to be processed falls within the area of S ₁ . Is a process of increasing the number of p by 1 when ▲ a ⁽⁰⁾ _n ▼ is not a reject (denoted by the? Symbol). Is a process for determining whether p is larger than P ₁ . The processing from to is the detailed processing of the processing in FIG. Is a process of registering p as a new P ₁ and N as the number of characters I to be stored in the row storage area # 1.
Is a process of storing the recognized line in the line storage area # 1.

FIG. 17 shows a detailed flowchart of the digit alignment process of FIG. In the flowchart of FIG. 17,
It is initializing with. Is a process in which the end of the contents stored in the row storage area is indicated in advance by a special large value X _M. X _M is set to a value larger than the value obtained by adding the value of W described later to X ₄ in FIG. 15 (a). Is paying attention,
The coordinates of the i-th character in row storage area # 1 ▲ x ⁽¹⁾ _i ▼, the coordinates of the j-th character in row storage area # 2 ▲ x ⁽²⁾ _j ▼, and the k-th position in row storage area # 3 This is a process for obtaining the minimum value Xmin of the coordinates ▲ x ⁽³⁾ _K ▼ of the character. In as long as the same as the X _M used in the processing of Xmin, it has been determined that the digit registration process of all of the characters is finished. W used in the above indicates the width of the x-coordinate that can be determined to be in the same digit. Now If is satisfied, ▲ x ⁽¹⁾ _i ▼ performs the processing when it enters the digit currently being processed, that is, the processing of. Is a process for registering the i-th character ▲ a ⁽¹⁾ _i ▼ in the line storage area # 1 as the character ▲ b ⁽¹⁾ _l ▼ after digit alignment. Is a process for increasing the value of i since the i-th process is completed. On the other hand, if ▲ x ⁽¹⁾ _i ▼ does not enter the digit currently being processed, the process is performed. The line # indicates that the line storage area # 1 has not been obtained (missing character) at the corresponding digit.
⁽¹⁾ It is a process of registering in _l ▼ (# has the same meaning as a space between characters in FIG. 5B). In the same manner as, the processes for row storage area # 2 and row storage area # 3 are performed. Is Xm as the position after digit alignment
This is the process of registering in to y _l . Is a process of increasing the number of characters l after digit alignment by 1. Is a process of registering the finally obtained number of characters l-1 as a variable L.

FIG. 18 (a) is a detailed flowchart (1) of the comprehensive process of FIG. Is a process of initializing a variable m for counting the characters in the total line. This is a repetitive process for repetitively performing the process from (a) to (d) for the number of digits L after digit alignment. The coordinate y _l a l-th digit of interest is determined whether or not to enter the area of S _1, the process proceeds to when entering the area of S _1. In row storage # 1
If the character ▲ b ⁽¹⁾ _l ▼ after the digit alignment is not # (if there is no character missing), it is registered as the character c _m of the general line. In, y _l is determined whether or not to enter the area of S _2, the process proceeds to when entering the area of S _2.
In line storage area # 2, the character ▲ b after column alignment
^{(2) If} _l ▼ is not # (if there is no character missing), it is registered as the character c _m of the comprehensive line.
If y _l is in the region of S ₃ , then
In line storage area # 3, characters after column alignment ▲ b ⁽³⁾
_{If l} ▼ is not # (if there is no missing character), it is registered as the character _cm of the general line.
Is a process of registering the number of characters m−1 of the total line in the variable M.

FIG. 18 (a) the row storage # 1 used for each of the regions S _1, S _2, S _3, # 2, is a process in the case of dividing the # 3, only the line storage used for each region A processing example that does not
It is shown in Fig. 18 (b). In FIG. 18 (b), a process of registering the digit number L after digit alignment as the character number M of the comprehensive line. Is a repetitive process for repetitively performing the process starting with.

In Fig. 18 (b), ▲ b ⁽¹⁾ _l ▼ is missing characters (when the # symbol is set in the digit alignment process).
But if it's not rejected (indicated by a? Sign),
Set ▲ b ⁽¹⁾ _l ▼ to c _l . ▲ b ⁽¹⁾ _{l If} ▼ is missing or rejected, go to. In ▲ b ⁽²⁾ When _l ▼ is neither rejected in character missing are with ▲ b ⁽²⁾ _l ▼ a c _l. ▲ b ⁽²⁾ _{l If} ▼ is missing or rejected, go to. In ▲ b ⁽³⁾ When _l ▼ is neither rejected in character missing are with ▲ b ⁽³⁾ _l ▼ a c _l.
▲ b ⁽³⁾ _{l If} ▼ is missing or rejected, proceed to.
After all, ▲ b ⁽¹⁾ _l ▼ and ▲ b ⁽²⁾ _l ▼ and ▲ b ⁽³⁾ _l ▼ are rejected or missing characters.In this case, set the reject to c _l . ing.

It should be noted that the initialization process of FIG.
The variables of J, K, P ₁ , P ₂ and P ₃ may be set to 0.

In the above embodiments, the case where the number of regions set in the field of view is 3 has been shown, but an example in which regions other than 3 are set is also possible if there are multiple regions. Also, when multiple areas are set in the field of view, it is not necessary to set them so that they do not overlap each other (so that they do not overlap). For example, as shown in FIG. 19, five areas are set while partially overlapping each other. Examples are possible.

〔The invention's effect〕

By using the present invention, it is possible to realize an optical character reader capable of reading an entire line even at a skew angle such that the entire line does not enter the field of view at once. To put it the other way around, this is an optical character reader according to the prior art (Fig. 20 (a)).
This means that the height of the field of view can be reduced compared to.
Figure 20 (b)). Even in a scanner having a visual field height shown in FIG. 20 (b), by using the present invention, the entire row can be read by applying the scanner to the row and moving the row up and down. If the height of the field of view can be reduced, the following effects can be obtained.

-Since the number of pixels required for the image sensor is reduced, a cheaper image sensor can be used and the cost of the device can be reduced.

-Since the field of view height is reduced, the size of the scanner is reduced and the operability of the scanner is increased.

-Since the visual field height is small, it is easy to reduce unevenness when the illumination light source in the scanner illuminates the document. Therefore, the design and development are easier than the illumination system of the optical character reader according to the related art.

Thus, the ripple effect of the present invention is great.

[Brief description of drawings]

FIG. 1 is a structural example of an optical character reader using the present invention,
FIG. 2 is an optical character reading device according to the prior art, FIG. 3 is an explanatory view up to a single character cutout process, FIG. 4 is an explanatory view of a single character cutout method, FIG. 5 is an explanatory view of digit alignment / majority result processing, Figures 6, 8 and 9 are illustrations of row movements in the field of view and the results of majority voting, and Figure 7 is an illustration of reading price tags,
FIG. 10 is an explanatory view of the problems of the prior art, FIG. 11 is an explanatory view of the operation of the present invention, FIG. 12 is an embodiment diagram of the row detecting means and the row interchange detecting means, and FIG. FIG. 14 is a schematic flowchart of the processing of the microprocessor, FIG.
Figures are explanatory diagrams of constants / variables, Figures 16 to 18 are detailed flow charts of line detection processing, digit alignment processing, and line total processing, respectively. Figure 19 is an explanatory diagram of how to set areas, and Figure 20.
The figure is an explanatory view of the field of view of an optical character reading apparatus using the present invention. 1 ... Scanner, 2 ... Hand, 3 ... Original, 4 ... Light source, 5 ... Lens system, 6 ... Image sensor, 7 ... Control binarization circuit, 8 ... Screen memory, 9 ... Single digit cutting circuit , 10 …… Single digit memory 11 …… Single character cutout circuit 12 …… Single character memory, 13 …… Character recognition circuit 14, 15, 16 …… Recognition result buffer 17, 17 ′ …… Digit matching processing unit 18 …… Majority processing Section 19 …… Format check section 20,130 …… Timer 21 …… Previous register, 22 …… Comparator 23 …… Output control section 30 …… Image sensor field of view 31,32,33 …… Line detection means 34,35 , 36 ... Line storage means 37 ... Digit alignment means 38 ... Line total processing means 39 ... Line exchange detection means 101, 106, 111, 116 ... Contents of line storage means # 1 102, 107, 112, 117 ... Contents of line storage means # 2 103, 108, 113, 118 ... Contents of line storage means # 3 104, 109, 114, 119 ... Digit alignment Result 105, 110, 115, 120 …… Total row result 121, 122, 123 …… Horizontal OR circuit 124, 125, 126 …… Black length determination unit 127 …… OR gate 128 …… Inverter gate 131 …… Microprocessor 132 ... ROM, 133 ... RAM.

Claims (7)

(57) [Claims] 1. An optical character reading apparatus for reading a character on an original by holding an image sensor housing (scanner) with a hand and applying it to the original, and a plurality of characters / symbols (hereinafter, simply referred to as characters). ) Is the target of recognition, an image sensor that fits at least one line of characters within the field of view, character recognition means that recognizes each character on the screen captured by the image sensor and outputs its position, and sets it in the field of view. A plurality of line detection means for detecting whether or not a part of the line exists in each of the plurality of displayed areas, and the recognition obtained from the character recognition means when the corresponding line detection means detects the line. A plurality of line storage means for storing lines; a digit alignment processing means for associating the recognized lines stored in each of the line storage means with each other based on the positional information of the characters; Based on the result of, the line detection means has a line totaling means for obtaining a totalized line in which the recognized lines stored in the line storage means are integrated, and a line replacement detecting means for detecting the replacement of the lines in the visual field. When the means detects the presence of a line, the line storage means corresponding to the line detection means stores the recognition line obtained from the character recognition means, and the digit alignment processing means stores the characters of the recognition line stored in the line storage means. An optical system characterized in that the lines are associated with each other by the line aligning means, the total line is obtained by the line totalizing means, and when the line exchange detecting means detects the line exchange, the contents of the line storing means are erased. Character reader. 2. The optical character reading device according to claim 1, wherein the line interchange detecting means determines that the line is interchanged when all the line detecting means detect no line. And an optical character reader. 3. The optical character reading device according to claim 1 or 2, wherein a black pixel is laterally arranged in a predetermined range provided in the visual field of the image sensor by the line detection means. It consists of a horizontal OR circuit that calculates the logical sum, and a black length determination unit that determines that the operation result of the horizontal OR circuit continues for a predetermined range in the vertical direction. An optical character reading device characterized by determining that a row exists when the calculation result of the horizontal OR circuit is detected. 4. The optical character reader according to claim 1 or 2, wherein the line detection means checks the position information of each character of the recognition line obtained from the character recognition result, It is a means to determine that a line exists by having a non-unrecognizable character having position information in a predetermined range, and a non-unrecognizable character having position information in the range in a recognition line already stored in the line storage means. Than, when many non-recognizable characters are obtained in the range,
An optical character reader characterized in that a newly obtained recognition line is stored in a line storage means. 5. The optical character reading device according to claim 1 or 4, wherein the line replacement detecting means determines that the recognized line is recognized when the character recognizing means repeatedly performs recognition processing. An optical character reading device comprising a time measuring means for measuring the time when the recognition line is not obtained, and determining the replacement of the line when the time when the recognition line is not obtained exceeds a predetermined time. 6. The optical character reader according to any one of claims 1 to 5, wherein the height of the visual field of the image sensor is 1 to 3 times the height of the character.
An optical character reading device characterized in that the position information of a character used for digit alignment is a position information in the horizontal direction within the visual field of the character within a range of about double. 7. The optical character reader according to any one of claims 1 to 5, wherein the height of the field of view of the image sensor is smaller than the line spacing of the character lines written on the document. An optical character reading device characterized in that position information used for digit alignment is position information in the lateral direction within the visual field of the character.