JPH01258558A - Method for deciding threshold density for original reader - Google Patents

Abstract

PURPOSE:To obtain a good binary picture which does not have the scratchiness of a line, etc., by fixing the threshold concentration of binarization based on the density distribution of a signal after shading correcting. CONSTITUTION:At a first original scanning, a density distribution measuring circuit 11 to which a multilevel digital signal is supplied from a shading correcting circuit 10 counts the number at every density value and stores it to a memory provided in the density distribution measuring circuit 11. When a first original scanning ends, density histogram data stored in the memory in the density distribution measuring circuit 11 are referred to from a CPU 13, and the upper limit value of the distribution area of the density of a background part is stored in a memory 14 as the threshold density of binarization. Picture data obtained at a second original scanning are compared with the threshold density stored in the memory 14 and binarized.

Description

[Detailed Description of the Invention] [Technical Field] The present invention is used in a document reading device used as an input device for a character recognition device, a figure/drawing recognition device, etc. Concerning a method for determining threshold concentration.

[Prior art]

An optical system that exposes and scans the image of the document on the document table, a lens that forms an image of the reflected light, a line image sensor that photoelectrically converts the image, and an A/D conversion circuit that converts this analog signal into a digital signal. , a shading correction circuit that performs shading correction on this multivalued digital signal, and a binarization circuit that compares the signal after shading correction with a threshold density, binarizes it, and outputs it to an external device. Conventionally, in a document reading device having a multi-level image, the automatic threshold density determination method when binarizing a multivalued image is as follows. A method is known in which the density value of the valley between the two peaks of the background part and the text/figure part in the density histogram is used as the threshold value. However, using method (■), for example, when the number of pixels in the text/figure parts is considerably smaller than the total number of pixels, such as in mechanical drawings, that is, when the mountains in the text/figure part are not noticeable in the density histogram, There was a problem that the threshold value could not be determined.

Figure 7 (al, (bl) is a diagram showing the situation, and in the figure, ■ and ■ represent the threshold concentration according to the methods of ■ and ■ mentioned above. That is, as shown in the same figure (b) In addition, the threshold value ■ cannot be determined unless the mountains in the text/figure parts are noticeable.

〔the purpose〕

The present invention solves the above-mentioned drawbacks of the conventional example, and even when the number of pixels in the character/graphic portion is relatively small compared to the overall number of pixels, such as a line drawing original, the quality of the original is good without blurring of the lines. It is an object of the present invention to provide a threshold value determination method for a document reading device that can obtain a binary image.

〔composition〕

In order to achieve this objective, the present invention measures the density distribution of a signal after shading correction, determines the upper limit density value of the background part from the measured density distribution, and uses it as the threshold density for binarization. It is characterized by

Embodiments of the present invention will be described below based on the drawings. 1st
The figure is a block diagram of a document reading device to which the present invention is applied. This device determines a threshold density for binarization in the first scan of the document, and uses that threshold value in the second scan. The configuration is such that the document is binarized during the first scan and output to an external device.

To explain its structure, a document table (transparent glass) 1 that supports a document 0 is fixed to the upper surface of the main body (not shown). An optical system consisting of an exposure lamp 2 and a mirror 3.4°5 is placed on the document table 1.
By reciprocating in the direction of arrow e along the lower surface of the lens, exposure scanning is performed during the forward movement. In this case, mirror 4.5 moves at half the speed of mirror 3 so as to maintain the optical path length. The light reflected from the original O by the scanning of the optical system, that is, the light reflected from the original 0 by the light irradiation of the exposure lamp 2, is reflected by the mirrors 3, 4.5, passes through the lens 6, and passes through the line image sensor. (COD) 7, and an image of the original O is formed on the line image sensor 7.

This line image sensor 7 converts the formed optical image into a corresponding electrical signal and sends it via an amplifier 8 ^
/D conversion circuit 9. This A/D conversion circuit 9 is
The signal supplied from the line image sensor 7 via the amplifier 8 is converted into a multivalued digital signal and output to the shading correction circuit 10. This shading correction circuit 10 performs shading correction on the multivalued digital signal supplied from the A/D converter 9, and sends it to the density distribution measurement circuit 11 during the first original scanning, and to the density distribution measurement circuit 11 during the second original scanning. A multivalued digital signal is supplied to the binarization circuit 12.

In the first document scanning, the shading correction circuit 10
The density distribution measuring circuit 11 supplied with the more multivalued digital signal counts the number of each density value and stores it in a memory provided within the density distribution measuring circuit 11. 1st
At the time when the first document scan is completed, the density distribution measurement circuit 11
The density histogram data stored in the memory of C
Referring from PUl3, detect the density distribution area of the background part, and use the upper limit of the area as the threshold density for binarization,
The value is stored in memory 14.

In the second document scanning, the shading correction circuit 10
The binarization circuit 12 that has been supplied with the more multivalued digital signal passes through the CPU 13 and compares it with the threshold density for binarization obtained during the first scan of the document in the memory 14 to generate a binary signal. interface 1 and send the resulting binary digital signal to interface 1.
Supply to 5.

The interface 15 outputs this binary data to an external device. One scan cycle ends when the second document scan is completed.

Further, a driver 16 in the figure represents a lighting control device for an exposure lamp, and a driver 17 represents a control device for a pulse motor 18 that reciprocates an optical system consisting of an exposure lamp 2, a mirror 3, and a '4.5.

In the present invention, as shown in FIG. 2, when a density histogram is viewed from the minimum density value to the maximum density value, the upper limit value of the density of the background portion is detected and used as the threshold density.

By employing such a threshold density determination method, it was possible to obtain a good binary image without blurred line drawings. Furthermore, this threshold value determining method is effective not only when the ratio of the number of pixels in a text/graphic portion to the total number of pixels is small, but also when the ratio is large (see FIG. 3).

To give a specific example of a detection method, as shown in FIG. When the number of pixels exceeds 2, the histogram peak begins, and when the number of pixels becomes less than 2, the histogram peak ends, and the density value at this time is taken as the threshold density for binarization.

In addition, as shown in Figure 5, a difference histogram is created from the density histogram, and when the difference histogram is viewed from the minimum to the maximum density value, the ends of the positive and negative peaks are detected and the background A method of determining the end of a portion may also be considered.

To give a concrete example of the detection method, as shown in Fig. 7, the difference value is looked at from the minimum density value to the maximum density value in the difference histogram, and when it reaches a certain value and becomes 1 or more, it is corrected. The beginning of the mountain, if it becomes L2 or less, the end of the positive mountain, and then the beginning of the negative mountain, if it becomes L3.
After that, when the value becomes 54 or more, it is considered as the end of the negative mountain, and the density value at this time is taken as the threshold density for binarization.

Although the above two methods are effective when used alone, by using them together, the accuracy of the threshold value was improved.

In addition, in the embodiment, one scan/scan is performed by scanning the document twice.
An example of a cycle was shown, but the following examples are: 1. Determine the threshold value when moving the optical system in the past, and output it as a binary value when returning. 2. Prepare an image memory for one document and store the scanned image data in it. It is also possible to make one document scan into one scan cycle by storing the information in the memory, taking it out after determining the threshold value, converting it into a binary value, and outputting it.

〔effect〕

As described above, according to the method for determining the binarization threshold density of the present invention, regardless of the ratio of the number of pixels in text/figure parts to the total number of pixels, it is possible to prevent blurring of lines, etc., faithful to the line drawing original. It is possible to obtain a good binary image with a small number of images.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a document reading device to which a threshold density setting method according to the present invention is applied, and FIG. FIG. 3 is a diagram for explaining a threshold concentration determination method according to an embodiment of the present invention, FIG. 4 is a diagram showing a specific example thereof, and FIG. 5 is a diagram for explaining a threshold concentration determination method according to another embodiment. A diagram for explaining the determination method, FIG. 6 is a diagram showing a specific example, and FIG. 7 (a
) and (b) are diagrams for explaining a threshold concentration determination method according to a conventional example. 0...Original, 1...Original table, 2. 3. 4. 5
... Optical system, 6... Lens, 7... Line image sensor, 9... A/D converter, 10... Shading correction circuit, 12... Binarization circuit.

Claims (1)

[Claims] An optical system that exposes and scans the image of the document on the document table, a lens that forms an image of the reflected light, a line image sensor that photoelectrically converts the image, and an A/D conversion circuit that converts this analog signal into a digital signal. , a shading correction circuit that performs shading correction on this multivalued digital signal, and a binarization circuit that compares the signal after shading correction with a threshold density, binarizes it, and outputs it to an external device. In the method for determining the threshold density of a document reading device having
A method for determining a threshold density for a document reading device, characterized in that an upper limit density value of a background portion is determined from a measured density distribution and is used as a threshold density for binarization.