JPH05298482A - Character reader - Google Patents

Abstract

PURPOSE:To provide a character reader where a character is normally and stably detected and segmented for a character picture generated by dots. CONSTITUTION:A photoelectric converting part 1 converts the character picture generated by the dots into an electric signal by optical scanning. An A/D converting part 2 converts the output signal of the photoelectric converting part 1 into a digital signal and stores it in the multi-picture memory 3 as the multi-character picture. A max. density value pick-up part 4 repeatedly scans the character picture which is stored in the multi-picture memory 3 so that max. density value data on respective scanning lines is picked-up. A row detecting and segmenting part 5 detects and segments a character row from the picked-up max. density value data. A character detecting and segmenting part 6 obtains the detecting and segmenting position of the character from the continuity of width between the maximal values of max. density value data on the respective scanning lines in the scanning direction of the character picture and detects and segments respective character units from the detected and segmented character row. A character recognizing part 7 recognizes the respective detected and segmented characters.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character reading device for optically reading characters and the like on a recording medium, and in particular,
The present invention relates to a character detection / cutout technique for detecting / cutting out character by character from an input multivalued character image.

[0002]

2. Description of the Related Art In a character reading device of this type, after converting an input multi-valued character image into a binary character image,
It is general to detect and cut out a character line and then to detect and cut out each character. In this case, in order to obtain a stable character image, the background must be uniform and the density difference between the background and the character must be clear.

Therefore, in the current optical character reading apparatus, it is premised that characters are printed or written in black on a white background with a uniform background, and the background of the object to be read may fluctuate due to uneven illumination or the like. As for the fluctuation of the background, a method for easily binarizing the character and the background is applied by shading correction or image enhancement using a spatial filter.

However, if the density level is locally reduced, especially when the periphery of the character is missing, the center of the character is displaced, which affects the recognition of the character. To avoid these,
A plurality of binarized images are extracted, each of them is subjected to character recognition, and the one with the best recognition result is the final result.

[0005]

[Problems to be solved by the invention] Shading correction,
Both the spatial filter can absorb a gentle background fluctuation as a background level, but it is difficult to remove a noise image having a signal level similar to a character by conventional image processing. In addition, the plurality of binarization processes for avoiding the loss of character images require raster scanning of the screen a plurality of times, which may unnecessarily increase the processing time or treat a part of the background as a character image. There is.

On the other hand, when characters are formed by dots, since each dot has a characteristic in character construction, the variation of the density value due to the imprinting method is great for the detection and cutout of characters and the recognition of characters. It will affect you. In particular, a decrease in the level of the density of the dots arranged in the periphery causes the image to be lost during binarization, and the identification image is greatly deformed. Therefore,
SUMMARY OF THE INVENTION It is an object of the present invention to provide a character reading device capable of always detecting and cutting out a character in a character image formed by dots.

[0007]

SUMMARY OF THE INVENTION A character reading apparatus of the present invention comprises a photoelectric conversion means for converting a character image formed by dots into an electric signal by optical scanning, and an output signal of the photoelectric conversion means as a digital signal. The maximum density for extracting the maximum density value data on each scanning line by repeatedly scanning the conversion means for converting into a multivalued character image and the multivalued character image converted by this conversion means. A value extracting means, a peak position detecting means for detecting the maximum value of the maximum density value data extracted by the maximum density value extracting means and a position thereof, respectively, and a length between the maximum values detected by the peak position detecting means. Is within a predetermined range, and if the result of comparison by the peak interval comparison means is within a predetermined range, the start point position and the end point of the maximum value within the continuous effective predetermined range are determined. Position detection means for detecting each position, character detection cutout means for detecting and cutting out character by character using the start point position and end point position detected by this position detection means as character detection cutout positions, and this character detection cutout The character recognition means for recognizing each character detected and cut out by the means.

Further, the character reading apparatus of the present invention includes a photoelectric conversion means for converting a character image formed by dots into an electric signal by optical scanning, and an output signal of the photoelectric conversion means for converting it into a digital signal. Conversion means for converting the multi-valued character image, and maximum density value extraction means for extracting the maximum density value data on each scanning line by repeatedly scanning the multi-valued character image converted by the conversion means. , A peak position detecting means for detecting the maximum value and the position of the maximum density value data extracted by the maximum density value extracting means, and a plurality of different lengths between the maximum values detected by the peak position detecting means. The peak interval comparison means for comparing whether or not it is within a predetermined range, and the result of the comparison by the peak interval comparison means is that, if it is within a plurality of predetermined ranges, the maximum value in the continuous effective predetermined range is determined. Position detecting means for detecting the point position and the end point position, and selecting means for selecting the start point position and the end point position closest to the specified width of the character from the respective start point position and end point position detected by the position detecting means. , Character detection cutout means for detecting and cutting out character by character using the start point position and end point position selected by this selection means as character detection cutout positions, and each character detected and cut out by this character detection cutout means Character recognition means for

[0009]

By detecting the character detection / cutout position from the continuity of the width between the maximum values of the maximum density values on each scanning line in the scanning direction of the character image, the character image formed by dots is always stable. It is possible to detect and cut out the characters that have been deleted.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

In FIG. 1, a photoelectric conversion unit 1 is, for example, a light source that irradiates a recording medium on which characters or the like are recorded with light, and a CCD type that receives reflected light from the recording medium and converts it into an electric signal. It consists of a photoelectric converter such as a sensor and an optical system that guides the light from the light source onto the recording medium and guides the reflected light to the photoelectric converter. Optical scanning on the recording medium converts it into an electrical signal (analog signal). It is designed to be converted.

The analog signal from the photoelectric conversion unit 1 is A /
The multi-valued digital signal is converted by the D conversion unit 2 and stored in the multi-valued image memory 3 as a multi-valued character image. The maximum density value extraction unit 4 repeatedly raster-scans the character image stored in the multi-valued image memory 3 in the same direction as the character line, so that the maximum density value data on each scanning line is extracted.

Next, in the line detection / cutout section 5, a character line is detected and cut out from the maximum density value data extracted as described above, and then in the character detection / cutout section 6, the detected and cutout character is extracted. Each character is detected and cut out from the line. Then, the character recognition unit 7 performs recognition (similarity calculation, etc.) for each detected and cut out character unit, and outputs the recognition result.

FIG. 2 shows the structure of the character detection / cutout section 6, and a method for detecting and cutting out dot characters from the maximum density value data will be described below with reference to this figure. First,
The maximum density value extraction unit 4 repeatedly raster scans the dot character image stored in the multi-valued image memory 3 to extract maximum density value data on each scanning line. An example of the extracted maximum density value data is shown in FIG. 3B, and FIG. 3A shows an example of the dot character image stored in the multi-valued image memory 3.

Next, as shown in FIG. 4, the background level calculator 11 sets Gmin to be the average value from the minimum density values in N division units, and Gmax to be the maximum density value of the entire data.
Then, the background level Gx is obtained as follows. The background level Gx is a threshold value for reducing the influence of background noise. Further, m is a constant, and is “8” here, for example. Gx = Gmin + (Gmax-Gmin) / m

Next, the subtracter 12 has a maximum density value extraction unit 4
The background level Gx calculated by the background level calculation unit 11 is subtracted from the maximum density value data extracted in step 3, and the calculation result is transferred to the peak position detection unit 13.

The peak position detector 13 detects a maximum value from the output data of the subtractor 12 and detects the position of the maximum value. As for the position detection of the maximum value, as shown in FIG. 5, when the density value satisfies Fi-1 <Fi> = Fi + 1, the Fi position is detected.

The peak-to-peak width calculation unit 14 calculates the width between the current maximum value position and the next maximum value position, and sends the calculation result to the peak-to-peak width comparison units 15 and 16. Peak width comparison section 1
Reference numerals 5 and 16 respectively compare the width value from the peak-to-peak width calculation unit 14 with preset first and second set values. Here, for example, the first set value is a permissible range between 1 dots, and the second set value is a permissible range between 2 dots.

As a result of this comparison, if the peak-to-peak width is valid, the first peak position (starting point position) and the last peak position (ending point position) of the section are the maximum values detected by the peak position detecting section 13. From the first character position memory 17 and the second character position memory 18 respectively.

Next, the overlapping position detecting unit 19 uses the starting point position and the ending point position stored in the first character position memory 17 and the second character position memory 18 to overlap the first character position memory 18 with the first character position memory 18. The starting point position and the ending point position of the character position memory 17 are extracted, and the more character-like one is extracted based on the judgment of both. FIG. 6 shows an example of duplication from the first character position memory 17 and the second character position memory 18. 6A shows an example in which both are the same, FIG. 6B shows an example in which the central dot is not detected, and FIG. 6C shows an example in which noise exists between characters.

The overlapping position detecting section 19 detects the starting point position and the ending point position of the first character position memory 17 which overlaps the second character position memory 18, and the character width setting value preset by the character width evaluating section 20. Compared to. When there are a plurality of first character position memories 17, the start point position and the end point position that are closest to the character width among the plurality of first character position memories 17 are the only candidates in the first character position memory 17. Then, which of the first character position memory 17 and the second character position memory 18 is closest to the character width setting value is evaluated.

As a result of this evaluation, the first character position memory 17
Is closest, a plurality of or a single overlapping start point position and end point position stored in the first character position memory 17 are stored in the character position determination memory 24 via the AND circuit 21 and the OR circuit 23. On the other hand, the second character position memory 18
Is closest, the single start point position and single end point position stored in the second character position memory 18 are stored in the character position determination memory 24 via the AND circuit 22 and the OR circuit 23. If both have the same width, the first character position memory 1
The start position and end position of No. 7 are stored in the character position determination memory 24. For example, the contents of the first character position memory 17 in the case of FIG. 6A, the contents of the second character position memory 18 in the case of FIG. 6B, and the first character position in the case of FIG. 6C. The contents of the memory 17 are stored in the character position confirmation memory 24, respectively.
Stored in.

When the start point position and the end point position are stored in the character position fixing memory 24 in this manner, each character is detected from the character line detected and cut out by the line detection / cutout unit 5 by using the start point position and the end point position as the final detection and cutting position of the character. The unit is detected and cut out and sent to the character recognition unit 7.

[0024]

As described above in detail, according to the present invention, the detected cutout position of a character is obtained from the continuity of the width between the maximum values of the maximum density values on each scanning line in the scanning direction of the character image. , For the character image formed by dots,
It is possible to provide a character reading device capable of always detecting and cutting out a stable character.

[Brief description of drawings]

FIG. 1 is a block diagram schematically showing an overall configuration of a character reading device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing in detail the configuration of a character detection / cutout unit.

FIG. 3 is a diagram showing an example of a multi-valued character image and maximum density value data for it.

FIG. 4 is a diagram for explaining calculation of a background level.

FIG. 5 is a diagram for explaining position detection of a maximum value.

FIG. 6 is a diagram for explaining the confirmation of a character detection cutout position.

[Explanation of symbols]

1 ... Photoelectric conversion unit, 2 ... A / D conversion unit, 3 ... Multi-valued image memory, 4 ... Maximum density value extraction unit, 5 ... Line detection cutout unit, 6 ... Character detection cutout unit , 7 ... Character recognition part, 13 ...
... Peak position detection unit, 14 ... Peak width calculation unit, 1
5, 16 ... Peak-to-peak width comparison unit, 17, 18 ... Character position memory, 19 ... Overlap position detection unit, 20 ... Character width evaluation unit, 24 ... Character position determination memory.

Claims (2)

[Claims] 1. A photoelectric conversion means for converting a character image formed by dots into an electric signal by optical scanning, and a conversion for converting an output signal of the photoelectric conversion means into a digital signal to obtain a multi-valued character image. Means and a maximum density value extraction means for extracting the maximum density value data on each scanning line by repeatedly scanning the multivalued character image converted by the conversion means, and the maximum density value extraction means. Peak position detecting means for detecting the maximum value of the extracted maximum density value data and its position, respectively, and a peak interval for comparing whether the length between the maximum values detected by the peak position detecting means is within a predetermined range or not. As a result of comparison by the comparing means and the peak interval comparing means, if the result is within a predetermined range, a position detecting means for detecting the start point position and the end point position of the maximum value within the continuous effective predetermined range, respectively. , Character detection cut-out means for detecting and cutting out character by character using the start point position and end point position detected by this position detection means as character detection cut-out positions, and each character detected and cut out by this character detection cut-out means. A character reading device comprising: a character recognition unit for recognizing. 2. A photoelectric conversion means for converting a character image formed by dots into an electric signal by optical scanning, and a conversion for converting an output signal of the photoelectric conversion means into a digital signal into a multi-valued character image. Means and a maximum density value extraction means for extracting the maximum density value data on each scanning line by repeatedly scanning the multivalued character image converted by the conversion means, and the maximum density value extraction means. The peak position detecting means for detecting the maximum value and the position of the extracted maximum density value data respectively, and comparing whether the lengths between the maximum values detected by the peak position detecting means are within a plurality of different predetermined ranges or not. If the result of the comparison between the peak interval comparing means and the peak interval comparing means is within a plurality of predetermined ranges, the start point position and the end point position of the maximum value in each of the continuous effective predetermined ranges are respectively determined. The position detecting means for detecting, the selecting means for selecting the starting point position and the ending point position closest to the specified width of the character from the respective starting point position and the ending point position detected by the position detecting means, and the selecting means for selecting the position. The character detecting and cutting means for detecting and cutting character by character by using the start point position and the end point position as the character detecting and cutting position, and the character recognizing means for recognizing each character detected and cut by the character detecting and cutting means. A character reading device characterized in that