JPS61145694A - Image reading system - Google Patents

Abstract

PURPOSE:To execute an image processing with a high recognizing ratio by storing a binary signal and a multi-value signal and a threshold signal which go to the base, at an image memory. CONSTITUTION:A register 51 is used as a means to store image information to an image memory 61, a multi-value signal and a threshold and the binary signal which go to be a base to generate a binary signal as unit picture element information are stored as a pair. For example, in eight bits of the register 51, a multi-value signal, an output of an A/D converting device 2, is stored in four bits, a threshold signal of an output of a threshold setting circuit 4 is stored at three bits and a binary signal of an output of a comparing circuit 3 is stored in one bit. The contents of the register 51 are written in an address #1 as information of 1 of the memory 61. In the same manner hereinafter, the information of a picture element 2 is written in an address #2, and the information of a picture element 3 is written in an address #3. Consequently, the information of one picture element is written in one address. Thus, the recognizing part to recognize an image by using the information of the memory 61, can execute the picture processing with a high recognizing ratio.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image reading system, and particularly to an image reading system that improves the reading accuracy of an optical reading device.

Image information such as characters and line drawings recorded on media is optical,
Through electrical manipulation, it is stored in the image memory as electrical image information. A character image is extracted based on this image memory, recognized as a character, and later displayed on, for example, a display device or printed on a printing device.

Therefore, the image information stored in the image memory is required to have high accuracy because it is necessary to identify the characteristics of characters, line drawings, and the like.

[Conventional technology]

FIG. 2 is a block diagram of main parts of a conventional image reading device.

The photoelectric converter 1 includes a plurality of photoelectric conversion elements arranged in a line, and the light receiving area of each photoelectric conversion element is, for example, 0. IX
o, about 1 mm", and the detected pixel unit is 0
．． I Xo, 1 n+n+".

The photoelectric converter 1 outputs different voltages for each pixel depending on the density of image information on the recording medium. This output voltage is converted into a digital signal by an analog/digital conversion circuit 2 and input to a comparison circuit 3.

In this comparison circuit 3, the value of the multilevel signal is compared with the value of the threshold signal, and if it is above the threshold, the output signal of the comparison circuit 3 becomes "1", and if it is below the threshold, it becomes 0.

This binary pixel signal is input to, for example, an 8-bit bumper register 5, and the binary signal of the pixel is written into the image memory 6 via the bumper register 5.

For example, pixels of an image recorded as shown in Fig. 4,
■...■... is the address of image memory 6! 1
1.112. 8 bits each are stored in $13...

[Problem that the invention seeks to solve]

Binary signal image information written to memory using conventional methods is sliced by a threshold value, so in the case of a character image with a characteristic such as a gradual decline in density, the characteristic is It cannot be expressed as .

For example, when the character ``hon'' recorded as shown in Figure 3(a) is binarized using a predetermined threshold because the horizontal - stroke at the bottom of the character is blurred, only the dark part appears in the image. The blurred parts are ignored. Therefore, when this image information is recognized by the subsequent recognition unit, the third
As shown in Figure (b), the result is that it is misidentified as a "tree".

Therefore, there is a problem in that the image information stored in the image memory using the conventional method is insufficient for recognition processing.

[Means for solving problems]

In order to solve the above problem, means is provided for storing in a memory a binary signal of a unit pixel constituting an image, the digital signal that is the basis for creating this binary signal, and the threshold signal as a set. This problem is solved by the image reading method of the present invention.

[Effect]

According to this invention, since binary signals, multi-value signals, and threshold signals are stored in the image memory as image information for each pixel, the recognition unit can detect areas with low density where the presence of images cannot be identified using binary signals. Characteristics of the characters can be identified by referring to the relationships among the three signals, and the recognition performance of the image reading device is improved.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention.

Note that the same reference numerals refer to the same objects throughout the plan.

Therefore, the configurations, operations, interconnections, input and output signals of the photoelectric conversion circuit 1, analog/digital conversion circuit 2, comparison circuit 3, and threshold value setting circuit 4 are the same as in the conventional example, and their explanations will be omitted.

As a means for storing image information in the image memory in this invention, an 8-bit register 51 is used to store a multi-value signal, a threshold signal, and its binary signal, which are the basis for creating a binary signal as unit pixel information. Store as a set. For example, 4 bits out of 8 bits of the register 51 are used to store the multi-value signal output from the analog/digital conversion circuit 2, 3 bits are used to store the threshold signal output from the threshold setting circuit 4, and 1 bit is used to store the output from the comparator circuit 3. Stores binary signals.

The contents of this register 51 are written into the image memory 61 of 8 bits per address as information on pixel 2 at address #1. Similarly, the information for pixel ■ is the address $12.
Then, the information of pixel ■ is written to address #3. Therefore, information for one pixel is written to one address.

In the above explanation, a character image is used as the image, and the number of bits of each signal is allocated assuming that the image information is stored in an 8-bit image memory. There is no problem in associating the bit and node distribution of each signal based on the above.

In addition, in the comparison circuit, a binary signal was created from the multi-value signal and the threshold signal, but the multi-value signal and the threshold signal are stored in the image memory, and the binary signal is generated from the two signals when performing recognition processing. A value signal may also be created.

Moreover, although the binary signal, multi-value signal, and threshold signal are stored in the same memory in the above-mentioned image memory, these three signals may be stored in separate memories.

〔Effect of the invention〕

As described above, according to the present invention, the image memory stores the binary signal, the multi-value signal that is the basis of the binary signal, and the threshold signal. The unit can perform image processing with a high recognition rate.

In addition, since each pixel is assigned an address, when cutting out and recognizing character images, it is possible to control addresses in memory, that is, rather than storing only binary signals that require consideration of bit positions. Since calculation processing becomes extremely easy, an image reading device with high processing capacity can be constructed.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the image reading method of the present invention, FIG. 2 is a block diagram of a conventional storage method, FIG. 3(a) is an example of a character image, and FIG. 3(b) is an example of a character image. An example of a recognized image, FIG. 4 is a schematic diagram of the relationship between character images and pixels. In the figure, 1 is a photoelectric converter, 2 is an analog/digital conversion circuit, 3 is a comparison circuit, 4 is a threshold value setting circuit, 5.5
1 represents a register, and 6.61 represents an image memory. Figure 1

Claims (1)

[Claims] A photoelectric converter that receives light from an image recorded on a recording medium, an analog/digital converter that converts the output signal of the photoelectric converter into a digital signal in units of pixels constituting the image, and a threshold value setting circuit. An image reading device comprising a comparison circuit that binarizes the digital signal according to an output threshold value, and a memory that stores the binary signal output from the comparison circuit, the binary signal being stored in the memory;
An image reading method characterized by storing a digital signal and a threshold value.