JPH0731716B2 - Optical character reader - Google Patents

Description

TECHNICAL FIELD The present invention relates to an optical character reading device for reading characters by scanning a sheet with a handheld scanner.

<Prior art and its problems> In supermarkets and department stores, POS (Point of Sales) systems that collect sales information for individual items and manage inventory are in widespread use. Handheld optical character readers are often used in this POS system.

FIG. 5 is a block configuration diagram of an example of a conventional optical character reading device, which holds a scanner 1 in a hand 2 and identifies a character while moving horizontally on a paper 3 on which a character is written. Is represented. The paper 3 is illuminated by the lamp 4, and the character pattern on the paper 3 is imaged through the lens system 5 on the image sensor 6 in which the photoelectric conversion elements are arranged in a planar shape (two-dimensional). Since the reflected light from the background area of the paper 3 and the reflected light from the character area are different from each other, the signals obtained by the photoelectric conversion elements corresponding to them are added to the control and binarization circuit 7, and the level judgment is performed. The determination of white and black binarization is performed. For example, the binarization determination result is output with the signal corresponding to the background area of the paper 3, that is, white being "0", and the signal corresponding to the character area, that is, black being "1". The signal “0”, which is binarized by the control and binarization circuit 7,
“1” is stored in the buffer register 8.

Here, a method of reading characters will be described. The photoelectric conversion element of the image sensor 6 will be referred to as a cell, and an area in which data is compressed to a resolution necessary for recognition will be referred to as a unit area. The character pattern on the paper 3 is the image sensor 6
If an example of the image formed above is shown in FIG. 6, FIG. 6 shows an example of the correspondence between each cell and the binarized signal output from the control and binarization circuit 7. The pattern composed of is called a screen. The control and binarization circuit 7 is the top row L ₁
Of B ₁ sequence B ₂ from row, B _3, ... B _n outputs a signal corresponding to the column, then, B ₁ of the next line _{L 2, B 2, B 3} , ... B n columns, then L ₃
The scanning of one screen is completed by outputting the signals corresponding to the B ₁ , ... B _n columns of the row and finally the L _m row. Subsequent vertical extracting circuit 9 of the buffer register 8 in FIG. 5 is a circuit for detecting a character area in the vertical direction, from the _first column B for each row
Check whether or not there is a "black" cell by the _Bn column, and check for "black"
If the number of rows in which cells exist is continuously in a certain range (determined from the size of the character), it is regarded as a vertical character area. In FIG. 6, lines L _j to L _{j + N-1 are} lines in which “black” exists continuously, and the vertical character area signal VA indicating this range is supplied to the horizontal cutout circuit 16 and the discrimination circuit 13. Output. The horizontal cutout circuit 16 is a circuit for determining whether or not a character has reached a predetermined position in the horizontal direction. For example, in the vertical character area (L _j line to L _{j + N-1} line), B ₁ When all the columns are “white” and part of the B ₂ column is “black”, the horizontal cutout signal VH indicating that the character has reached the predetermined position is output to the discrimination circuit 13. Discrimination circuit 13, when it is determined that the horizontal position is correct, a circuit for identifying the character vertical character area (e.g., L _{j + N-1} line from L _j line) as a target, a row feature extraction circuit
It has a partial feature extraction circuit 11 and an identification processing circuit 12. The line feature extraction circuit 10 extracts and classifies the features C _i of each line in the detected character area. The partial feature extraction circuit 11 performs feature classification based on a plurality of signals of C _i sent from the row feature extraction circuit 10 and outputs D _i . The discrimination processing circuit 12 discriminates characters from the order of the signals D _i sent from the partial feature extraction circuit 11, and outputs the result D _out . More specifically, the feature extraction circuits 10 and 11 classify the features of the row according to the length of consecutive "black", that is, "1" and the number of consecutive occurrences obtained in each row. For example, in the character pattern of FIG. 6, and L _j row It will be classified as follows. Next, the features classified in order from the upper row are combined and identified. For example L _j , This combination narrows the character identification range, and then L _{j + 3}
Line characteristics, below If the features are sequentially combined up to and the range is narrowed down to "2"
Can be identified as

In this way, the scanner 1 is laterally moved by the hand 2 and scanned, whereby the characters are read one by one.

By the way, in the above-mentioned conventional technique shown in FIG. 5, it is difficult to correctly move the scanner 1 in the horizontal direction, so that it is difficult to read it correctly, and since it is operated by the hand 2, the operability is poor. .

On the other hand, in order to improve the operability, a wide area image sensor is used so that multiple characters / symbols to be read can be placed within the field of view, and reading is performed simply by placing the scanner on the characters / symbols without moving. An optical character reader was also developed. FIG. 7 shows a conventional optical character reader of this type.

In FIG. 7, reference numeral 21 is a scanner, which is used to read characters, symbols and the like written on the paper 23 by only placing it on the paper 23 with the hand 22. The paper 23 is, for example, a price tag in which the upper side in the POS system is described. Reference numeral 24 is a light source, 25 is a lens system, and 26 is an image sensor, and has a field of view of information described on the paper 23, that is, almost the same size as the price tag. 27 is a control and binarization circuit, which converts the analog signal which is the output signal of the sensor 26 into a binarized signal, and sends it to the memory 28. The memory 28 stores binarized data of almost the entire field of view of the sensor 26.

FIG. 8 shows an explanation of the binarized data of the image sensor 26. It is an image sensor 26 of p × q pixels, and paper 23
Is the entire field of view of the information description part. In FIG. 8, lines for distinguishing each pixel and part of the image are omitted.

Characters and symbols are identified by the identification circuit 30.
Since 30 is for identifying one character at a time, one character's worth of data is retrieved from the memory 28. The control circuit 29 takes out data corresponding to m × n pixels, which is the processing capacity of the identification circuit 30, from the memory 28 and stores it in the buffer register 8.

In FIG. 8, first, the data from the row L _{1 to the} row L _{m of the} columns B ₁ to B _n are fetched from the memory 28 and transferred to the buffer register 8.
The processing of the character lines after the buffer register 8 is the same as the conventional one. Next, the data from the L ₁ row to the L _m row of the B ₁ +1 to B _n +1 columns, that is, the data moved to one pixel side is stored in the memory 28.
To the buffer register 8. Similarly from B ₁ +2
Transfer the to L _m rows from L ₁ row of B _n +2 columns, which was Kukaee, transfers the data from the L ₁ row to B _q rows L _m rows.

By this data transfer, the same scanning as moving the conventional scanner 1 with the hand 2 is performed. When the L ₁ row to B _q string up L _m rows is finished, turn shifts the line down. For example,
If the field of view from line L ₁ to line L _m is about twice the height of the character, it is shifted by about m / 2 pixels, that is, from line L _{m / 2 to} line L _{3 m / 2} , starting from column B ₁ Transfer data up to column B _n, and repeat the same operation. By the way, in the prior art shown in FIG.
Data corresponding to one character is extracted from the entire field of view for identification processing, and then the data is moved laterally pixel by pixel to extract data for one character for identification processing. For this reason, the same portion of the pixel is repeatedly identified, and it takes a long time to identify the entire field of view.

In view of the above-mentioned conventional technique, the present invention includes most of the characters, symbols, etc. written on the paper only by placing the scanner on the paper by hand, and the operability is good and high-speed processing is possible. An object of the present invention is to provide a handheld optical character reader.

<Means for Solving Problems> The optical character reading device of the present application for solving the problems has the following features. That is, an image sensor having a field of view of a plurality of characters or symbols, a means for storing the output signal of the image sensor as digitized image data, and one character or symbol is completely stored in the stored image data. In the optical character reading device for recognizing a character or a symbol for one character from the image data of the extracted small area, means for extracting the image data of the small area,
A means for detecting the starting end of a character or a symbol for each character in the small area by ORing the pixels in the vertical or horizontal direction of the image data of the small area; and the detection from a predetermined position in the small area. The means for detecting the distance to the starting end and the means for collectively moving the position of the small area by the detected distance one character at a time are used to sequentially cut out characters or symbols one by one. It is a character reader.

<Example> An example of the present invention is shown in FIG. 21 is a scanner,
Just put it on the paper 23 with your hand 22, and the letters written on the paper 23,
It is for reading symbols and the like. The paper 23 is, for example, a price tag in which information in the POS system is described. Reference numeral 24 is a light source, 25 is a lens system, and 26 is an image sensor, which has a field of view substantially the same size as the information recording surface of the paper 23, that is, the price tag. Reference numeral 27 denotes a control and binarization circuit, which converts an analog signal which is an output signal of the image sensor 26 into a binarized signal and sends it to the memory 28. The memory 28 stores binarized data of almost the entire field of view of the image sensor 26.

Characters and symbols are identified by the identification circuit 30.
Since 30 is used to identify each character, small area image data for one character is taken out from the memory 28. The control circuit 31 takes out the small area image data corresponding to m × n pixels which is the processing capacity of the identification circuit 30 from the memory 28, and the buffer register 8
It is provided with a small area extracting means for storing in.

Further, the vertical cutout circuit 9 detects a vertical character area, the horizontal cutout circuit 16 determines whether a character or the like has reached a predetermined horizontal position, and the line feature extraction circuit 10 detects it. circuitry for extracting distinguishing features C _i of each line in a character area, partial feature extraction circuit 11 circuit which outputs a D _i performs further characterized partitioned from the characteristic C _i, identification processing circuit 12 is the signal D _i
It is a circuit that identifies characters from the order of and outputs the result D _out .

Now, in FIG. 1, a vertical cutting circuit 9 and a horizontal cutting circuit 1
The small area image data from the buffer register 8 is input to the distance detection circuit 32 as well as 6 and the row feature extraction circuit 10. The distance detection circuit 32 uses the m × n pixel data of a small area in which one character is completely contained, and determines the number of pixels from a predetermined position of this small area to the start end of each character or symbol in the same area. The number of pixels SLN is detected and the number of pixels SLN up to m × n pixel data, which is the data for the next character, is output. This will be explained with reference to FIG. 2 using the cut-out screen. First, as shown in FIG.
When the image data of a small area of m × n pixels, in which the symbol portion is completely contained, is taken out, the identification circuit 30 cuts out the character in the horizontal direction and is equivalent to the distance from the predetermined position A of the small area to the start end of the character. The number of pixels l ₁ is detected and l ₁ is output to the control circuit 31 as a signal SLN. In this case, since the output VA of the vertical cutout circuit 9 is input to the distance detection circuit 32, in the pixel data to the distance detection circuit 32, the character after vertical cutout exists in the upper and lower approximately center.

The output SLN of the distance detection circuit 32 is sent to the control circuit 31,
The control circuit 31 shifts the screen by the number of pixels l ₁ based on SLN, then extracts m × n pixel data for one character and stores it in the buffer register 8. This state is shown in FIG. 2 (b).

Then, in FIG. 2 (b), similarly to the above, the number of pixels l ₂ from the predetermined position A of the small area to the adjacent character is detected (as a result, the distance between the two characters is detected). , Signal SL
When l ₂ as N is output from the distance detection circuit 32 to the control circuit 31, the control circuit shifts the screen as shown in FIG. In this way, the number of pixels up to the adjacent character is detected, and the screens for that pixel are collectively shifted to accelerate the extraction of data for one character. That is, the control circuit 31 also functions as a means for moving the position of the small area.

FIG. 3 shows an example of the configuration of the distance detection circuit 32, which is composed of an OR circuit 33, a shift register 34, and a control ROM 35. Binary data of each column is sequentially input to the OR circuit 33 for each row on the screen, and in this OR circuit 33, the OR with the output of the shift register 34 of n stages corresponding to the horizontal row and n columns of the screen. This is taken again and stored in the shift register 34 again, and the logical sum in the vertical direction is taken for each column in all rows. That is, the column in which the pixel that becomes the character is first found is found. The result of OR is ROM35 for control
The control signal corresponding to the number of pixels is created and read in the ROM 35. The ROM 35 also has a function of detecting and removing noises such as those smaller than the characters, and the program also determines whether or not there is noise. In case of noise, SLN does not output.

In this way, since the start of a character can be detected in the horizontal direction of the screen by the number of pixels, the screen can be shifted to that position all together. In the description of FIG. 2, the case of horizontally shifting the pixels has been described. However, the pixels may be vertically shifted as shown in FIG. 4, and the vertical pixel number l ₃ is detected in FIG. 4 (a). However, it is also possible to take out the data by shifting the number of pixels as shown in FIG. In the description of FIG. 2, the case where the pixels are viewed from the left to the right has been described as an example. However, even when the pixels are viewed from the right to the left, if the above description is reversed left and right, it is exactly the same. It is clear that you can do it.

<Effects of the Invention> The present invention uses a scanner having a sensor having almost the entire field of view for describing information on a sheet such as a price tag, and by simply placing the scanner on the sheet, the characters or the like described on the sheet can be read. It is possible to realize an optical reading device with extremely good operability.
Moreover, when the character-by-character identification processing is performed, the high-speed processing can be performed by detecting the start edge of the character and processing the data in the character area.

[Brief description of drawings]

1 to 4 show an embodiment of the present invention, FIG. 1 is a configuration diagram, FIGS. 2 (a), (b) and (c) are state diagrams showing a state where the screen is shifted, and FIG. 3 is a distance. Configuration diagram of an example of a detection circuit,
4 (a) and (b) are state diagrams showing a state where the screen is shifted,
FIGS. 5 to 8 are conventional examples, FIG. 5 is a configuration diagram of an example, FIG. 6 is an explanatory diagram of binarized data, FIG. 7 is a configuration diagram of another example, and FIG. It is an explanatory view of conversion data. In the figure, 1 ... Scanner, 2 ... Hand, 3 ... Paper, 4 ... Lamp,
5 ... Lens system, 6 ... Image sensor, 7 ... Control and binarization circuit, 8 ... Buffer register, 9 ... Vertical cutout circuit, 10 ... Row feature extraction circuit, 11 ... Partial feature extraction circuit, 12 ...
Identification circuit, 13 ... Identification circuit, 16 ... Horizontal cutout circuit, 21 ... Scanner, 22 ... Hand, 23 ... Paper, 24 ... Light source, 25 ... Lens system,
26 ... Image sensor, 27 ... Control and binarization circuit, 28 ...
Memory, 29 ... Control circuit, 30 ... Identification circuit, 31 ... Control circuit,
32 ... Distance detection circuit.

Claims (1)

[Claims] 1. An image sensor having a field of view of a plurality of characters or symbols, a means for storing output signals of the image sensor as digitized image data, and a character or a symbol for one character from the stored image data. In the optical character reader for recognizing a character or a symbol of one character from the extracted image data of the small area, a means for extracting the image data of the small area that completely fills in the vertical or horizontal direction of the image data of the small area. Means for detecting the starting end of each character or symbol in the small area by ORing the pixels in the direction, and means for detecting the distance from the predetermined position in the small area to the detected starting end, An optical character in which a character or a symbol is sequentially cut out character by character by providing a means for moving the position of the small area collectively for each character by the detected distance. Winding device.