JPH04181392A - Optical character reader - Google Patents

Abstract

PURPOSE:To securely read characters out of a document, etc., where colors other than a dropout color are used by displaying an image which is decided as a character by using a color indicating the character and then correcting the decision making result. CONSTITUTION:A binarizing part 80 performs a binarizing process so that a label image decided as a character by a label decision part 50 or a label image indicated with position information from an operation part 70 is effective and others are ineffective; and the obtained binarized image is sent out to a character recognition part 90 and a character recognizing process is performed. In this case, the label image decided as the character is displayed by using the color indicating the character and the decision result is corrected according to the displayed image. Consequently, characters can be read out of general documents on the market and documents produced without the consideration of processing by an optical character reader.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to an optical character reading device for reading characters written on a form or the like.

(Prior art) Generally, optical character reading devices that read characters written on forms, etc. process black and white (binary) images in which black is valid (character data) and white is invalid (background). It is targeted for. Therefore, in a scanning unit that optically scans a document 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 processing target of optical character reading devices,
The color used to print text boxes and characters that do not need to be read on forms is a color that is determined to be white by an optical character reader (threshold logic in the scanning unit), that is, a dropout color. Colors called or used. This dropout color is determined depending on the wavelength of the light source used in the scanning section.

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

Incidentally, in general, many commercially available slips and forms that are not intended to be processed by an optical character reading device are used. Therefore, it is required that the characters written on these slips and forms be read by optical character reading devices.

(Problem to be solved by the invention) However, commercially available slips and forms that are not intended to be processed by optical character reading devices (dropout colors or not used) are available in a variety of colors. Because of this, even if an optical character reading device is used to read the characters, the characters to be read may drop out due to binarization, or both the background and the characters may be determined to be black. For this reason, it has been impossible to read characters recorded on forms other than OCR forms.

The present invention has been made in view of the above points, and provides an optical character reading device that can read characters on commercially available general slips and forms that are not intended to be processed by an optical character reading device. The purpose is to

[Structure of the Invention] (Means for Solving the Problems) The present invention includes a scanning means for detecting a color image on a sheet on which characters, etc. are written, and a color image detected by the operation means for at least one sheet. a storage means for storing, and a color image stored in the storage means;
a label image detection means for detecting a label image consisting of a set of connected pixels for each of a plurality of colors; and a label for determining whether or not a label image detected by the label image detection means is a label image of text. a determining means; a display means for changing only the label image determined to be a character label image by the label determining means to a specific color and displaying the color image detected by the scanning means;
human-powered means for manually inputting positional information indicating an arbitrary position in the image displayed by the displaying means; and generating a binary image based on the positional information input by the inputting means and the determination result by the label determining means. Binarization means,
The apparatus further comprises a character recognition processing means for performing character recognition processing based on the binary image obtained by the binarization means.

(Function) According to such a configuration, a label image that is determined to be a character is displayed in a specific color indicating the character, and the determination is corrected based on this displayed image. can.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an optical character reading device according to the same embodiment. In FIG. 1, reference numeral 10 denotes a control section, which controls the entire operation of the apparatus. The following sections are controlled by this control section 1.0. 20 in the diagram
1 is a scanning unit that 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 the color image output from the scanning section 20, for example, for one form. Reference numeral 40 in the figure is a labeling unit, which labels the color image stored in the sheet memory 30 by pixel connectivity for each color and stores the image (hereinafter referred to as a label image). In the figure, 50 is a label determining section, which determines whether the image labeled by the labeling section 40 is an image or a character. Reference numeral 55 in the figure represents a character string extracting unit, which extracts character strings from the array of label small images determined to be characters by the label determining unit 50.

In the figure, 60 is a display section, which displays images stored in the labeling section 40 or the sheet memory 30 under the control of the control section 90. 70 in the figure is an operation unit, and any one in the image displayed on the display unit 60
A point can be specified, and position information indicating the specified position is output to the control section 10.

80 in the figure is a binarization unit, which enables (black, black,
For example, “1”) and disable the others (white, for example “0”).
”) and outputs the obtained binary image to the character recognition unit 90. The character recognition unit 90
Character recognition processing is performed based on the binary image input from the digitization section 80.

Next, the operation of this embodiment will be explained.

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 examines the connectivity of pixels for each of the plurality of colors in the color image stored in the sheet memory 30, labels each connected set, and creates an image for each label (label image). Store. When the images for each color are labeled, the label determination unit 50 determines whether or not each color represents a character based on the color and connectivity of the label image for each color, for example. As a method for character determination, the label image is thickened by a predetermined value, and if it becomes a continuous (clump) image, it is determined that it is a character portion. In other words, since the character image is fine, it takes advantage of the fact that the image is blurred by making it thicker. Alternatively, it is determined by the number of lines existing in a unit area.

That is, in the case of characters, the fact that there are more lines than others is utilized.

The control unit 10 causes the display unit 60 to display the human-powered image.

At this time, the color of the label image determined to be text by the label determination unit FPO is converted to, for example, black and displayed. However, if an image showing characters and a background adjacent to the same image have the same color, the brightness or color of the background is changed and displayed so that the characters and the background can be distinguished.

Note that the color in which images determined to be characters are displayed is not limited to black.

Here, an arbitrary point can be specified on the image displayed on the display section 60 by, for example, moving a mouse cursor on the operation section 70. In other words, it is determined whether the label image determined to be a character is correct or not, and if a change is necessary, an arbitrary point in the area to be changed in the displayed image is specified, and then that area is determined as a character. Specify whether to use it as a background.

The position instruction from the operation unit 70 is notified to the control unit 10. Note that at this time, it may be possible to designate a plurality of areas. Thereby, even if the label determination unit 50 makes an incorrect determination, it can be easily corrected.

The control unit IO outputs the position information and label information from the operation unit 70 to the label determination unit 50. The label determination unit 50
The label image indicated by the position information is read out from the labeling unit 40, and the character determination process is performed again.

The character string extraction unit 55 extracts character strings from the array of label images determined to be characters by the label determination unit 50. The character string extraction section 55 converts the extracted character string image into a binarization section 8.
Transfer to 0. The binarization unit 80 converts the label image determined to be a character string by the label determination unit 50 or the operation unit 70
Binarization is performed so that the label image including the positional information (indicating a character) inputted from is made valid (for example, "1"), and the other label images are made invalid (for example, "0"). 2
The digitization section 80 outputs the image obtained by the binarization to the character recognition section 90.

The character recognition unit 90 extracts 1 from the binary image from the binarization unit 80.
A character pattern is cut out for each character and character recognition processing is performed.

In this way, a label image determined to be a character can be displayed with a different color, allowing the operator to indicate whether or not the determination is incorrect. Therefore, since the image to be subjected to character recognition processing is reliable, it is possible to reliably read characters even on forms and the like that are not intended for optical character reading devices.

[Effects of the Invention] As described above, according to the present invention, an image determined to be a character can be displayed in a color indicating the character, and the determination can be corrected. This makes it possible to reliably read characters on forms and the like that use colors other than dropout colors, which are not intended for processing by an optical character reading device.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an optical character reading device according to an embodiment of the present invention. DESCRIPTION OF SYMBOLS 10... Control part, 20... Scanning part, 30... Sheet memory (storage means), 40... Labeling part (label image detection means), 50... Label determination part, 55...
- Character string extraction unit, 60...Display unit, 70...Operation unit (input means), 8o...Binarization unit, 90...Character recognition unit. Applicant's agent, patent attorney Takehiko Suzue! s1 diagram

Claims (1)

[Scope of Claims] A scanning means for detecting a color image of a sheet on which characters or the like are written, a storage means for storing at least one sheet of color image detected by the operation means, and a storage means for storing the color image detected by the operation means for at least one sheet; a label image detection means for detecting a label image consisting of a set of connected pixels for each of a plurality of colors from the stored color images; and a label image detected by the label image detection means, which is a label image of a character. a display means for displaying the color image detected by the scanning means by changing only the label image determined to be a text label image by the label determining means to a specific color; an input means for inputting positional information indicating an arbitrary position in the image displayed by the display means, and a binary image based on the positional information input by the input means and a determination result by the label determination means. An optical character reading device comprising: a binarization means for generating a binary image; and a character recognition processing means for performing character recognition processing based on the binary image obtained by the binarization means.