JPS5990175A - Binary coding circuit - Google Patents

Abstract

PURPOSE:To obtain the correct threshold value by sampling the quantized video signals in a fixed range for a fixed period of time and deciding the threshold value of those quantized video signals in response to the state of the frequency distribution of the video signal. CONSTITUTION:A timing control circuit transmits the quantized video signal to a histogram calculating circuit as soon as the scan of the 1st screen is started. The histogram calculating circuit produces a density histogram in real time in response to the scan after counting the frequency of each picture element for each quantized density level. The timing control circuit stops the production of the density histogram when the histogram calculating time elapses and indicates the start of execution for detection of the threshold value. A threshold value deciding circuit decides the threshold value at a time point T2 for the density histogram on the basis of the threshold value deciding algorism that decides a valley V1 between the background density and the desired picture density as the threshold value. This decided threshold value is held until the end of the scan of the corresponding screen.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a binarization circuit that receives an input image and separates a target image from a background or the like.

[Background technology and its problems]

With the development of pattern recognition technology, it has become possible to input various images into a computer and extract, recognize, and judge the features of the images. When performing this Bakun'd recognition using a computer, the original image is divided into a two-dimensional array of picture elements (1), and then an appropriate quantization level ( After the image is divided into two, it is common to perform various types of image processing.

Furthermore, during image processing, it is often necessary to perform processing to detect whether or not a target image exists in the original image. The target image generally has a light absorption history or light transmittance that is ``higher'' or ``lower'' than its surroundings, such as character recognition (2Oi ℃ is a character in the background, cell recognition). In this case, it is a cell.

Here, the common method for extracting the target image from the background is to compare the video signal of the original image with an appropriate threshold level and determine whether the video signal exceeds the threshold. is also called "binarization." This “binarization”
The original image subjected to is converted to a 0" or °゛1" array,
Based on this, specific logic (processing to detect the presence or absence of the target image from two or processing to extract the size of the target image by counting the number of picture elements that have become It I I+) is performed. As described above, "binarization" processing is one of the extremely important processes for image processing.

FIG. 1 shows an example in which a time-series analog video signal is subjected to "binarization" processing using three types of threshold levels VTHL + TH2 + VTH3. In FIG. 1, A is a graph of a video signal in which the vertical axis represents the value of the video signal and the horizontal axis represents time. B is the threshold value. This is a graph obtained by "binarizing" the video signal shown in graph A when the old version is used. Similarly, C is a graph when the threshold value is VTH2, and D is a graph when the threshold value is VTH2.
3, the video signal shown in graph A is "binarized"
This is the processed graph.

As can be seen from Figure 1, the size of the 1" part when "binarized" changes depending on the size of the threshold value, and therefore, the size of the 1" part when "binarized" is When determining the presence or absence or measuring the size of the target image,
This will lead to erroneous judgment.

Therefore, during the "binarization" process, the threshold level must be set correctly.

However, when inputting an image, there are usually changes in the light intensity of the light source, the force of the optical/imaging system, changes in brightness due to <%, changes in the photoelectric conversion element of the photoelectric conversion element, and changes in the video signal amplification system. Due to various causes such as the performance of gain, offset, etc., the video signal level always changes, and the accurate threshold value always changes accordingly. Conventionally, the objective and the threshold of the clear image were determined based on the previous screen.
A method of determining the value by analogy from the image take was adopted, but as mentioned above, the accurate 'ILllllil value always changes by C2, and the video signal level changes between the previous screen and the next screen. If so, it was Tabata to get the correct darkness value.

[Purpose of the invention]

The present invention has been made in view of the above circumstances, and when an image is input and the target image is separated from the background, an appropriate threshold value is set regardless of the type of image, and the video signal is correctly separated. The purpose of this invention is to provide a binarization circuit that converts into values.

[Summary of the invention]

In order to achieve the above object, the present invention samples a quantized video signal in a certain range for a certain period of time, creates a histogram to obtain the frequency distribution of the video signal, and determines the quantized video signal in a certain range according to the state of the frequency distribution. The method is characterized in that the dark value of the signal is determined and the quantized video signal is binarized every certain range (2) from the determined threshold value C2.

〔Example〕

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

In FIG. 2, a target object 2 is placed between a flash light source 1 and an optical/imaging system 3. A photoelectric conversion element 4 that accumulates an optical image of the subject 2 in the extension line of the light source 1, the subject 2, and the optical imaging system B, and converts the optical image into a video signal that is an electrical signal by scanning the accumulated image. is provided.

The electrical signal output from the photoelectric conversion element 4 is sent to two video amplification circuits 5 that amplify the video signal to an appropriate signal level.

The output amplified by the video amplification circuit 5 is input to an A-D converter B which samples and quantizes the input signal at appropriate timing. The output from the A-D converter 6 is sent to the gate circuit 7
The signal is input to a histogram calculation circuit 8 which counts the frequency of appearance at each level of the quantized video signal. The calculation result of the histogram calculation circuit 8 is the state of the histogram (
The signal is input to an open chain determining circuit 9 which determines a threshold value based on two steps. The threshold signal from the open chain determination circuit 9 and the quantized video signal from the A-D converter 6 are sent to a comparator circuit 10ζ, which compares the dark value and the quantized video signal at once and determines their magnitude. Then, a binarized video signal of "0" or "1" is output from the comparison circuit 10.The timing control circuit 11 controls the lighting of the light source 1, the scanning control of the photoelectric conversion element 4, and the A
-I) Sampling control of converter 5 and converter 6, control of gate circuit 7 through which the quantized video signal passes, control of the period of histogram calculation of threshold value determination circuit 9, control of start of calculation for threshold value determination, etc. are performed.

Next, the operation of the binarization circuit according to the present invention configured as described above will be explained.

The light source 1 emits light according to a command from the timing control circuit 11,
A subject 2 is illuminated, and its optical image is formed on the light-receiving surface of a photoelectric conversion element 4 via an optical/imaging thread 6. Photoelectric conversion element 4
Since it is an accumulation type, the optical image formed on the light-receiving surface (2) becomes an electrical image according to the change in the amount of charge or electrical conductivity, and the charge image is transferred to the light-receiving surface (2) until it is read out by scanning. Next, in response to the scanning signal from the timing control circuit 11, the photoelectric conversion element 4 sequentially reads out the charge images while shifting the position little by little while scanning in the one-dimensional direction in the case of a two-dimensional image, and converts the image signal into a video signal. The video signal of the photoelectric conversion element 4 is input to the video amplification circuit 51, where it is amplified to the level required for the next stage of processing.This amplified video signal is input to the A-D converter 6, where it is , receives the sampling signal from the timing control circuit 11, quantizes the video signal to the 1111th order, and converts it into discrete video information.Furthermore, at the start of scanning of the first screen, the timing control circuit 11 controls the gate circuit 7. The quantized video signal is sent to the histogram calculation circuit 8.The histogram calculation circuit 8 takes the quantization density level (the frequency of two occurrences) of each picture element as an argument, and creates a density histogram in real time according to the scan. After the histogram calculation time, the timing control circuit 11 starts the gate circuit 7.
to stop creating the concentration histogram. At this point, in the schematic diagram of Fig. 4, a density histogram of the video signal of the area Z in the upper diagonal area of one screen has been created, and the graph depends on the subject 2, but generally it is shown in Fig. 5. shows a curve like that shown below. In FIG. 4, H3 represents the horizontal scanning direction, Vs represents the vertical scanning direction, T1 represents the light source lighting time, T2 represents the threshold value determination time, HR represents the histogram calculation area, and DC represents the binarization area. In FIG. 5, the horizontal axis represents density levels, and the vertical axis represents the frequency of appearance of each density level. In FIG. 5, the density level P1 that gives one peak of appearance frequency is the background density that occupies most of the video signal, and the other peaks P2. P3 is one of the values indicating the density of the target image. Of course, if the desired image does not exist in one screen, the density histogram will be displayed as P2. Peaks such as P3 do not appear.

When the histogram calculating circuit 8 finishes creating the density histogram, the timing control circuit 11 instructs the threshold determining circuit 9C to start executing threshold detection based on the density histogram. The threshold value determination circuit 9 determines the threshold value at time T2 in FIG. 4 according to a threshold value determination algorithm that determines the valley 1 between the background density and the target image density in the density histogram as the threshold value, for example. The value is held until the scanning of the screen (DC> in FIG. The magnitude is compared, and if the quantized video signal is larger than the threshold, it outputs a signal IT i II, and if the quantized video signal is smaller than the threshold, it outputs a signal It D I.
Outputs a signal I.

Finish the "binarization" of the first screen as described above, and then convert the second screen in the same way.
, and before scanning that screen, the light source is turned on, the histogram is cleared, and in the area at the beginning of the screen scan, the video signal is based on C2, and the second screen is set separately from the threshold value of the first screen.
A threshold value for the screen is determined, the second screen is "binarized" using this threshold value, and thereafter, the same process as for the first both sides is repeated.

The operation timing of the above-mentioned binarization circuit will be explained with reference to FIG. A point light source 1 irradiates a first screen of a subject 2 with bulk X, and the optical body of the first screen is accumulated. The photoelectric conversion element 4 scans the first screen at X on the entire scanning range chart.

Next, the histogram calculation circuit of the histogram calculation circuit 8
- Determine the first value with X. The first screen is binarized using the X on the binarization chart of the comparison circuit 10.

[Modified example]

In the above embodiment, the light source was configured with a flash light source, but 1
If the variation in the amount of light is negligible during the completion of scanning the screen, a continuous light source such as an incandescent lamp may be used.

Further, in the above-described embodiment, the transmitted light of the object was converted into an image signal, but the same effect can be expected with the reflected light of the object.

In addition, in the above embodiment, a two-dimensional photoelectric conversion element was used, but a -dimensional photoelectric conversion element can also be used in the same manner as in the embodiment by using the video signal of the part where scanning is started. effect can be obtained.

Further, in the above-described embodiment, the histodaram division circuit 8,
Although the threshold value determination circuit 9, the comparison circuit 10, the gate circuit 7, etc. are configured by hardware, they can be realized by computer software having similar functions.

〔Effect of the invention〕

As described above, according to the present invention, when performing "binarization" of a screen, the necessary threshold value is determined and updated based on image data within a certain range, so that , an appropriate threshold value is always obtained.

Therefore, even if the video signal level fluctuates, such as fluctuations in the light intensity of the light source, brightness changes due to optical/imaging systems, or gain/offset fluctuations in the video amplification system (2), the video signal level (1) Since binarization is performed using a threshold value (2), a correct binarized signal can always be obtained.

[Brief explanation of the drawing]

Figure 1 is a graph showing examples of changes in the number M of binarization (digitalization) in which an analog video signal is binarized using various threshold levels.
The figure is a block diagram of a binarization circuit according to the present invention (2), Fig. 3 is an operation timing chart of the binarization circuit according to the present invention, and Fig. 4 is a diagram showing processing of one side by the binarization circuit according to the present invention. FIG. 5 is a graph showing an example of the S degree histogram when determining the most suitable country according to the present invention. 1...Light source, 2...Subject, 3... Optical imaging system, 4... Photoelectric conversion element, 5... Video amplification circuit,
6... A-D converter, 7... Gate circuit, 8
···histogram! i-1 arithmetic circuit, 9... Threshold value determination circuit, 10... Comparison circuit, 11... Timing control circuit. Agent Patent Attorney Nori Kensuke Chika (1 idiot) 3 made fishi he/°ha/

Claims (1)

[Claims] (11 light source, a photoelectric conversion element that accumulates a light source image of the subject from the irradiation C2 of the light source and converts the optical image into a video signal by scanning the accumulated image, and a photoelectric conversion element that amplifies the video signal of the photoelectric conversion element. a video amplification circuit; an A-D converter that quantizes the video signal of the video amplification circuit; a histogram calculation circuit that counts the frequency of appearance at each level of the quantized video signal from the A-D converter; The state of the frequency distribution of each level as an argument by the histogram calculation circuit (a threshold determination circuit that determines the threshold of the quantized video signal from two; and a comparison circuit that compares the quantized video signal with the quantized video signal and binarizes the quantized signal.