JP2925275B2 - Optical character reader - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial application field) The present invention relates to an optical character reader for reading characters and the like written on a form and the like.

(Prior Art) In general, an optical character reader that reads characters and the like written on a form or the like processes a black-and-white (binary) image in which black is valid (character data) and white is invalid (background). It is targeted for. Therefore, in a scanning unit that performs optical scanning on a form or the like to detect an image, binarization corresponding to white / black is performed based on a certain threshold logic.
For this reason, OCR, which is the object of processing by the optical character reader,
The colors used for printing character entry frames, characters that do not need to be read, and the like on a form are colors that are determined to be white by an optical character reading device (threshold logic in the scanning unit), that is, dropout colors. The called color is used. This dropout color is determined according to the wavelength of the light source used in the scanning unit.

Therefore, in the case of an OCR form, unnecessary color printing of a background or the like is dropped out, and only the written characters (character data) to be read are detected and subjected to character recognition processing.

By the way, in general, a lot of commercially available slips, forms that do not require processing by an optical character reader, and the like are used. For this reason, it is required that the characters and the like written on these slips and forms are read by an optical character reading device.

(Problems to be Solved by the Invention) However, commercially available slips, and slips or the like that do not assume the processing by the optical character reader (no dropout color is used) use various colors. Therefore, even if the reading is performed by the optical character reading device, the character to be read is dropped out by binarization, or both the background and the character are determined to be black. For this reason, characters recorded on forms other than the OCR form cannot be read.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and provides an optical character reading device capable of reading characters of a form or the like that is not premised on processing by a commercially available general slip or optical character reading device. The purpose is to do.

[Configuration of the Invention] (Means for Solving the Problems) The present invention provides a scanning unit for detecting a color image on a sheet on which characters and the like are written, and a storage unit for storing the color image detected by the scanning unit. A storage unit, a label image detection unit that detects a label image composed of a set of connected pixels for each color from the color image stored in the storage unit, and a label for each color detected by the label image detection unit A ruled line detecting means for detecting a ruled line image recorded on the sheet based on whether or not a condition that the image is a ruled line is displayed, and a display for displaying the ruled line image detected by the ruled line detecting means Means, input means for inputting position information indicating an arbitrary position in an area surrounded by the image of the ruled line displayed by the display means, and input means input by the input means. A label image of a specific color having the largest area among the label images for each color detected by the label image detecting means in the area indicated by the position information detected, and the detected label image of the specific color is detected. Means for generating a binary image in which the label image is invalid and other label images are effective; and the binarization means obtained by the binarization means.
And character recognition processing means for performing character recognition processing based on the value image.

(Operation) According to such a configuration, a label image of a background unnecessary for character reading is detected, and a binary image including a character pattern is generated based on the detected label image. Therefore, in a general optical character reading device, Characters can be read for forms and the like in colors that are not dropped out.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of an optical character reading apparatus according to the embodiment. In FIG. 1, reference numeral 10 denotes a control unit which controls the overall operation of the apparatus. The following units are controlled by the control unit 10. In the figure, a scanning unit 20 scans a form or the like to be processed as a color image and outputs a color image signal to the sheet memory 30. The sheet memory 30 stores, for example, one form of the color image output from the scanning unit 20. In the figure
Reference numeral 40 denotes a labeling unit that labels a color image stored in the sheet memory 30 according to the connectivity of pixels for each color and stores the image (hereinafter, referred to as a label image). In the figure, reference numeral 50 denotes a ruled line detection unit which determines and detects ruled lines from the label image stored in the labeling unit 40. In the figure, reference numeral 60 denotes a display unit, which is a labeling unit.
40, or an image stored in the sheet memory 30 under the control of the control unit 10. In the figure, reference numeral 70 denotes an operation unit which can designate an arbitrary point in the image displayed on the display unit 60 and outputs position information indicating the designated position to the control unit 10. In the figure, reference numeral 80 denotes a binarization unit, which inputs position information from the operation unit 70 from the control unit 10,
A labeling unit having the largest area in an area (ruled area) surrounded by a ruled line including a point indicated by the position information
The binarization is performed such that the label image stored in 40 is invalidated (white), and the resulting binary image is output to the character recognition unit 90. The character recognition section 90 performs a character recognition process based on the binary image input from the binarization section 80.

 Next, the operation of the embodiment will be described.

First, a form to be processed is scanned by the scanning unit 20 and detected as a color image. The detected color image is stored in the sheet memory 30.

Next, the labeling unit 40 checks the connectivity of pixels for each of a plurality of colors in the color image stored in the sheet memory 30, performs labeling for each connected set, and performs image processing for each label (label image). Is stored. When the labeling is performed on the image for each color, the ruled line detection unit 50 determines and detects the ruled line based on the label image of each color. As a result, information indicating that the image is a ruled line is added to the label image determined to be a ruled line. The determination / detection of a ruled line is performed, for example, by projecting for each label image and determining whether or not the condition (length, thickness, etc.) that the projection is a ruled line is satisfied.

The control unit 10 reads the image determined as a ruled line by the ruled line detection unit 50 from the sheet memory 30 or restores the image from the label image and displays the image on the display unit 60. Here, the operation unit 70
For example, by moving the mouse cursor, the display 60
In the image of the ruled line displayed in (1), any one point (that is, one area) in the enclosed area can be designated. The position instruction from the operation unit 70 is notified to the control unit 10. At this time, a plurality of areas may be designated.

The control unit 10 transfers the position information from the operation unit 70 to the binarization unit 80. The binarizing unit 80 reads out, from the labeling unit 40, a label image for each color in a closed region surrounded by a ruled line including the designated position information. Then, a label image having the largest area in the designated closed region is detected. That is, the label image detected here is an image based on the background color of the form. The binarizing unit 80 invalidates the detected label image (for example, “0”) and validates other label images (corresponding to other colors) (excluding the ruled line portion) in the closed area (for example, “1”). Binarization is performed as in "). As a result, an image including the character pattern can be obtained.

For example, as shown in FIG. 2, it is assumed that a ruled line is detected at a broken line position in the drawing by reading a form with a blue background, and a point in a closed area surrounded by a ruled line at the upper left is pointed. . In this case, in the designated closed area, the label image with the blue background is detected as the largest area. If a character is written in this area, the detected label image looks like a white character.
The binarization unit 80 binarizes the image by making the background part invalid and the character part valid, and outputs a binary image to the character recognition unit 90 as shown in FIG.

The character recognition unit 90 cuts out a character pattern from the binary image from the binarization unit 80 and performs a character recognition process.

When a plurality of closed areas are designated from the operation unit 70, the binarization processing by the binarization unit 80 and the character recognition unit
The character recognition processing by 90 is sequentially executed for the designated area.

In this way, the label image having the largest area in the closed area specified as the processing target is set as the background, and binarization is performed based on the label image. Also, the character recognition processing can be performed reliably.

[Effects of the Invention] As described above, according to the present invention, a label image having a maximum area in a designated closed region is detected as a background,
Since character recognition processing can be performed based on the remaining label images, characters such as general slips on the market and forms using colors other than dropout colors that are not assumed to be processed by an optical character reader It is possible to read.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an optical character reading apparatus according to one embodiment of the present invention, and FIGS. 2 and 3 are views for explaining processing in a binarizing unit. 10 control section, 20 scanning section, 30 sheet memory (storage means), 40 labeling section (label image detection means), 50 ...
Ruled line detection unit, 60 ... display unit, 70 ... operation unit (input means), 80
... Binarization unit, 90 ... Character recognition unit.

Claims (1)

(57) [Claims] 1. A scanning means for detecting a color image on a sheet on which characters and the like are written, a storage means for storing the color image detected by the scanning means, and a color image stored in the storage means A label image detecting means for detecting a label image composed of a set of connected pixels for each color; and whether or not a condition that the label image for each color detected by the label image detecting means is a ruled line is satisfied. , A ruled line detecting means for detecting a ruled line image recorded on the paper, a display means for displaying a ruled line image detected by the ruled line detecting means, and a ruled line image displayed by the display means Input means for inputting position information indicating an arbitrary position in an area surrounded by the area, and the laser in an area indicated by the position information input by the input means. A label image of a specific color having the largest area among the label images for each color detected by the image detection means, invalidating the detected label image of the specific color and validating other label images And a character recognition processing unit for performing a character recognition process based on the binary image obtained by the binarization unit. Optical character reader.