JPH03154183A - Optical character reader - Google Patents

Abstract

PURPOSE:To read character information described in an original without fail by extracting more than an allowable value, for which skew is generated from the tip of a sheet image by conveyance, as a white level. CONSTITUTION:A sheet 11 generating the skew during the conveyance is irradiated with light from a light source 12 and reflected light is made incident from the sheet 11 to a photoelectric conversion part 13, converted to multilevel image data and inputted to a white level memory 14 and a binarizing circuit 16. The multilevel image data stored at the prescribed address of the white level memory 14 are converted to the while level from the right end to L and from the left end to L in the tip of the image data. The binarizing circuit 16 compares the multilevel image data, which are inputted from the photoelectric conversion part 13, with the converted white level, binarizes the data in an image buffer 17 while maintaining an oblique state by the skew. Thus, all described characters can be read.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) The present invention relates to a device for recognizing character information written on a manuscript, particularly when the manuscript is conveyed in an oblique state. ,
The present invention relates to an optical character reading device that prevents a decrease in character reading rate in pattern recognition due to so-called skinny.

(Prior Art) In order to read characters written on a document, a conventional optical character reading device performs binarization based on the white level of the white background of the document being conveyed (hereinafter referred to as a sheet). I am using it. The above binarization process involves one scan of the white level of the leading edge of the sheet (clean area) obtained by irradiating light onto the sheet being conveyed and photoelectrically converting the reflected light from the sheet into memory. Memorize minutes. The white level stored in the memory is compared with the level of the extracted image of the standardized pattern, which is the image area of the sheet to be conveyed in advance. By comparison, the image area of the document with the above white level becomes the white value,
Characters and the like on the original are converted into black values and stored in the sheet buffer. In the binarization process described above, if the sheet is conveyed in an oblique state, it is difficult to accurately read the character information stored in the document.

(Problems to be Solved by the Invention) In the conventional apparatus, when a sheet is conveyed in a skinny state and stored in the image buffer 17, as shown in FIG.
The binarized sheet image 18 is stored in an oblique state. When this sheet image 18 is stored in an oblique state, the English letters rA written on the sheet image 18
Among J to rlJ, rGJ becomes unreadable, and the reliability of the device may be lowered due to the inability to accurately read the English characters on the sheet image 18, and a countermeasure for this problem has been desperately needed.

The present invention has been made in view of the above, and an object of the present invention is to accurately extract the image area of the document even when the document is conveyed at an angle, thereby extracting character information written on the document. An object of the present invention is to provide an optical character reading device that reliably improves the reliability of the reading device.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above object, the present invention provides that a document on which character information is written is conveyed, light is irradiated onto the document, and light is reflected from the document. Area detection means for detecting an image area of a document being conveyed from light; storage means for storing the image area of the document detected by the area detection means; and the leading edge of the image area of the document stored in the storage means. an extraction means for extracting as an image area of the document from both left and right ends of the document to both right and left ends of the leading edge of the image area of the document when the document is conveyed diagonally; The present invention further comprises: binarization means for comparing the image area of the document detected by the area detection means and binarizing the extracted image area.

(Function) In the optical character reading device having the above configuration, a document on which character information is written is conveyed, the document is irradiated with light, and the image area of the document is conveyed by light reflected from the document. Detect. The image area of the document to be detected extends from both left and right ends of the leading edge of the document stored in the storage means to both left and right ends of the leading edge of the image area of the document when the document is conveyed diagonally. Extract the image area. This extracted image area is compared with the detected image area of the document, and the extracted image area is binarized, so that character information written on the document can be accurately read.

(Example) Hereinafter, the present invention will be explained in detail using the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the optical character reading device of the present invention.

The optical character reading device includes a light source 12. Photoelectric conversion section 13
(area detection means) and a white level memory 14 (storage means).

The light source 12 is a document that is being transported (hereinafter referred to as a sheet).
For example, any device that irradiates light such as a fluorescent light may be used.

The photoelectric conversion unit 13 has a photoelectric conversion element and an A-D converter, and converts the reflected light incident from the sheet 11 into multivalued image data, which is a digital image signal, and converts it into a white level memory 14 and binarization, which will be described later. It is output to the circuit 16. The photoelectric conversion element is, for example, a semiconductor sensor such as a CCD type or MO3 type image sensor, or a vidicon.

There are image dissectors, etc.

The white level memory 14 stores image data input from the photoelectric conversion section 13 at a predetermined address.

Further, the optical character reading device has a white level correction circuit 1.
5 (extraction means), a binarization circuit 16 (binarization means), and an image buffer 17.

The white level correction circuit 15 reads the image data stored at a predetermined address in the white level memory 14 and stores it in advance in a memory (not shown) from both left and right ends of the leading edge (clean area) of the image data. The extraction pattern is compared with the extracted pattern, which is the standardized pattern of the above manuscript. By comparison, the white level is outputted from the white level memory 14 to the binarization circuit 16 as the white level with the distance Ll from the right end and the distance Ll from the left end of the leading edge of the image data as the image area of the document.

The binarization circuit 16 compares the multivalued image data inputted from the photoelectric conversion section 13 with the white level extracted by the white level correction circuit 15 and binarizes the data. That is, the binarization circuit 16 converts the English letters "A to ■" stored in the sheet 11 out of the white levels extracted by the white level correction circuit 15 into black values. In addition, the binarization circuit 1
6 sets the image data to a white value among the white levels other than the above-mentioned English characters.

The image buffer 17 stores the sheet image binarized by the binarization circuit 16 at a predetermined address.

The sheet image 18 from which the white level of the sheet 11 is extracted by the white level correction circuit 15 and binarized by the binarization circuit 16 will be described with reference to FIG.

A so-called skewed sheet image 18 in which the sheet 11 is oriented diagonally while being conveyed to a predetermined address in the white level memory 14 described above is stored diagonally on the right side in the figure.

When the image data is stored in the white level memory 14, the white level correction circuit 15 adjusts the stored sheet image 1.
The range Ll from the left end in the figure and Ll from the right end of the tip of 8 is extracted as the white level.

Here, the distance LL2 in the figure extracted as the white level is greater than or equal to the set value of the allowable skew range in which all the left and right ends of the skewed sheet image 18 are included. The white level extracted by the white level correction circuit 15 is compared with the multivalued image data manually inputted from the photoelectric conversion unit 13 by the binarization circuit 16, and the English letters rA-IJ written on the sheet 11 are The area of the other sheet image 18 is binarized into a black value and a white value.

The sheet image 18 binarized by the binarization circuit 16 is stored at a predetermined address in the sheet buffer 17. FIG. 3 shows the sheet image 18 stored at a predetermined address in the sheet buffer 17. Sheet image 1
8 are all stored in the sheet buffer 17, so compared to the conventional sheet buffer 1.
In comparison with the sheet image 18 stored in 7, it is possible to avoid the unreadability of the English letters rGJ.

Next, the operation of this embodiment will be explained.

First, after power is turned on to the apparatus, light is irradiated from the light source 12 onto the sheet 11 that has been skewed during conveyance, and reflected light from the sheet 11 is incident on the photoelectric conversion section 13 . The incident reflected light is converted into multivalued image data by the photoelectric conversion section 13 and input to the white level memory 14 and the binarization circuit 16. The multivalued image data stored at a predetermined address in the white level memory 14 is stored in the white level correction circuit 1.
5, the image data is converted into a white level from the right end to 1 and from the left end to Ll. The converted white level is compared with multi-valued image data manually inputted from the photoelectric conversion unit 13 by the binarization circuit 16 and is binarized.

In the binarized sheet image 18, the English characters on the sheet 11 are converted to black values, the areas of other sheet images 18 are converted to white values, and are stored in a predetermined address of the image buffer 17 in a diagonal state due to skew. Ru. The alphabetic characters in the sheet image 18 stored in the sheet buffer 17 are recognized by a recognition circuit (not shown).

As a result, all of the English characters rAJ to rGJ written on the sheet 11 conveyed in a skewed state can be read, so that the sheet 11 can be
It is possible to avoid a situation in which rGJ among the English characters written in is unreadable.

Further, as an application example of this embodiment, it is also possible to apply it to an image processing device that processes various images and a mark reading device that reads various marks (emblems).

As described above, the present invention can be implemented with various modifications within the scope of the invention.

[Effects of the Invention] As explained above, according to the present invention, since the white level is extracted from the leading edge of the sheet image to a value exceeding the allowable value for the skew caused by conveyance, even if the document is conveyed at an angle, the document By accurately extracting the image area, it is possible to reliably improve the reliability of the device that reads the character information written on the document.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the optical character reading device of the present invention, FIG. 2 is a diagram showing a binarized sheet image according to the present invention, and FIG. A diagram showing a sheet image stored in a buffer. FIG. 4 is a diagram showing a sheet image stored in a conventional sheet buffer. 12... Light source 13... Photoelectric conversion unit 14... White level memory 15... White level correction circuit 16... Binarization circuit 17... Sheet buffer 18... Sheet image

Claims (1)

[Scope of Claims] An area detecting means for detecting an image area of the transported original document by irradiating light onto the original document and detecting an image area of the transported original document based on light reflected from the original document; A storage means for storing an image area of a document detected by the detection means; and an image area of the document when the document is conveyed diagonally from both left and right ends of the leading edge of the image area of the document stored in the storage means. an extraction means for extracting the left and right ends of the leading edge of the document as an image area of the document, and comparing the image area extracted by the extraction means with the image area of the document detected by the area detection means, An optical character reading device comprising: binarization means for binarizing the extracted image area;