JPH07129753A - Method and device for setting optimum binarized threshold - Google Patents

Abstract

PURPOSE:To calculate and set an optimum binarized threshold following up the variance of the illumination condition. CONSTITUTION:Before the measurement processing in a visual recognition device, points P0 and P1 are obtained by an average density in a preliminarily determined monitor area and the optimum binarized threshold to determine a line L showing the relational formula between the density and the optimum binarized threshold with respect to the picture obtained in the illumination condition that illuminance is 0 and the picture obtained in the illuminance condition that illuminance capable of stable measurement processing is set. A control part measures the average density in the monitor area at intervals of a prescribed time, and, if the variance of the density is detected, the density is applied to the relational formula of the line L to calculate the optimum binarized threshold.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for binarizing a grayscale image obtained by picking up an object with a predetermined binarizing threshold value and performing a measuring process of a feature amount or the like. In particular, the present invention relates to an optimum binarization threshold value setting method and apparatus for calculating and setting an optimum binarization threshold value that follows changes in illumination conditions.

[0002]

2. Description of the Related Art In a visual recognition apparatus, for example, when an image of an object is taken by a television camera, if the illumination conditions change due to changes in the surrounding environment such as natural light or indoor lighting, the density of a grayscale image obtained by the image pickup changes. I have something to do. Due to this change in the illumination condition, the binary image obtained by binarizing the grayscale image is not always constant even for the same target object, which hinders the measurement process of the feature amount and the like.
Therefore, in order to always obtain a constant binary image regardless of the change of the illumination condition, it is necessary to correct the binarization threshold value by following the change of the illumination condition.

Conventionally, in order to correct the binarization threshold value, the average density of a predetermined monitoring region of an image obtained by picking up an object is measured, for example, at regular intervals to monitor variations in illumination conditions. However, when a change in the illumination condition is detected from the measured value of the density, the binarization threshold value is shifted by the amount of change in the pre-average density for correction. In FIG. 6, a straight line L indicates the conventional method of correcting the binarized threshold value. In the figure, the horizontal axis is the density and the vertical axis is the binarization threshold value. Since the straight line L has an inclination of 45 degrees, the binarization threshold shift amount ΔY matches the density variation amount ΔX. To do.

[0004]

However, since the density variation amount ΔX and the binarization threshold shift amount ΔY do not always have the same value, when the illumination condition varies, only the concentration variation amount ΔX becomes 2. Even if the binarization threshold value is shifted and corrected, the optimum binarization threshold value that follows the variation of the illumination condition cannot be obtained, and the binarized image generated by the binarization threshold value is the same object. However, it is not always constant.

The present invention has been made in view of the above problems, and an optimum binarization threshold value setting method capable of setting an optimum binarization threshold value that follows the fluctuations even if the illumination condition changes. The purpose is to provide the device.

[0006]

The invention according to claim 1 is a method for setting an optimum binarization threshold value that follows changes in illumination conditions, and is obtained by changing the illumination conditions and imaging. For each image, the density of a predetermined monitoring area is measured, and the optimum binarization threshold value under each illumination condition is determined to obtain the relationship between the density and the optimum binarization threshold value in advance. In the measurement process, when an image obtained by picking up an object is measured, the density of the monitoring area is measured to monitor the variation of the illumination condition, and the variation of the illumination condition is detected by the measured value of the concentration. The optimum binarization threshold value is calculated and set by applying the measured value of the density to the relationship between the density obtained in advance and the optimum binarization threshold value.

According to the second aspect of the present invention, a predetermined monitoring region is set for the image obtained under the illumination condition when the illuminance is zero and the image obtained under the illumination condition where the stable measurement process is possible. Is measured, the optimum binarization threshold value under each illumination condition is determined, and the relationship between the density and the optimum binarization threshold value is obtained in advance.

According to the third aspect of the present invention, the density of a predetermined monitoring region is measured for each image obtained under the illumination condition of different illuminance that enables stable measurement processing.
The optimum binarization threshold value under each illumination condition is determined, and the relationship between the density and the optimum binarization threshold value is obtained in advance.

According to a fourth aspect of the present invention, there is provided an apparatus for setting an optimum binarization threshold value that follows changes in illumination conditions, wherein each image obtained by capturing images under different illumination conditions is previously set. A storage unit that stores the relationship between the obtained density of the predetermined monitoring region and the optimum binarization threshold value, and an image obtained by capturing an image of an object during the measurement process,
Concentration measuring means for measuring the concentration of the monitoring area to monitor variations in illumination conditions, and the concentration stored in the storage means for the measured value of the concentration measured by the concentration measuring means and the optimum binarization threshold value. And a binarization threshold value setting unit for setting the optimum binarization threshold value calculated by the computation unit. is there.

[0010]

When a change in the illumination condition is detected, the measured value of the density of the predetermined monitoring area is applied to the relationship between the density obtained in advance and the optimum binarization threshold value.
Since the optimum binarization threshold value is calculated, it is possible to set the optimum binarization threshold value that follows changes in illumination conditions.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an example of the overall structure of a visual recognition device in which the present invention is implemented. In the illustrated example, the television camera 1
Captures an object under illumination by an illumination device (not shown) to generate a grayscale image, and outputs the video signal to the A / D converter 2. The A / D converter 2 converts an analog video signal into a digital grayscale image data.

The grayscale image data is output to the display controller 9 and the image memory 3, and the grayscale image is stored in the image memory 3. The density measuring unit 4 measures the average density of the grayscale images stored in the image memory 3 from the densities of all the pixels included in the predetermined monitoring area. Window data for setting the monitoring area is written in the window memory 5.

FIG. 2 shows a grayscale image 20 obtained by picking up an image of an object. In FIG. 2, 21 indicates an image part of the object and 22 indicates an image part of the background. In the background image portion 22, a window 23 defining the above-mentioned monitoring area is set.

Returning to FIG. 1, the binarization circuit 6 inputs the grayscale image data from the A / D converter 2 and binarizes it with a predetermined binarization threshold value to generate binary image data. . An LUT (look-up table) is used for the binarization circuit 6, and the table data is written by the CPU 12.

The other image memory 7 has a binary value from the binarization circuit 6.
Input the value image data and store the binary image. The measurement processing unit 8 measures, for the binary image stored in the image memory 7, a characteristic amount necessary for visual recognition such as an area and a peripheral length.

The display control unit 9 inputs grayscale image data, binary image data, window data, measurement results, etc.,
Either one of them is selected or combined, and the D / A converter 1
Output to 0. The D / A converter 10 converts the input data into an analog signal and outputs it to the video monitor 11. The video monitor 11 displays a grayscale image, a binary image, or the like on the screen according to the input analog signal.

The CPU 12 is a control body that controls each component in series, and calculates the relational expression between the density and the optimum binarization threshold value, which will be described later, and the optimum binarization threshold value using this relational expression. Calculate and set the. The CPU 12, the ROM 13, the RAM 14, the I / O via the CPU bus 17
An O port 15 or the like is connected, and an input device 16 such as a keyboard or a mouse is connected to the I / O port 15. The ROM 13 stores a program, and R
The AM 14 stores data such as coefficients that form a relational expression.

FIG. 3 shows two points P ₀ ,
A straight line L represents the relational expression calculated from P ₁ , that is, the relational expression between the concentration and the optimum binarization threshold value. On the other hand, the point P ₀ is the average density of the image in the above-mentioned monitoring area and the optimum binarization threshold value determined by the operator for the image obtained under the illumination condition when the illuminance is zero. It depends on the type of camera.
In this experiment, if a large number of objects having different concentrations are used for sampling, the accuracy of this point P ₀ is improved.

On the other hand, the point P ₁ is the average density in the above-mentioned monitoring area and the optimum binary value determined by the operator for the image obtained under the illumination condition of the illuminance that enables stable measurement processing. This is a point determined by the threshold value of the conversion and is obtained when the reference pattern required for recognition is registered using the model of the reference object. Assuming that the optimum binarization threshold value is y and the average density of the image in the monitoring area is x, the following equation (1) can be obtained from the two points P ₀ and P ₁ . In addition, a and b in the equation (1) are coefficients.

[0020]

[Equation 1]

FIG. 4 is a straight line L showing a relational expression calculated from two points P ₂ and P ₃ obtained by experiments by another method.

The points P ₂ and P ₃ are adjusted by the operator with the average density of the image in the above-mentioned monitoring area for each image obtained under the illumination condition of different illuminance that enables stable measurement processing. This is a point determined from the determined optimum binarization threshold value, and is similarly obtained when the reference pattern necessary for recognition is registered using the model of the reference object. Assuming that the optimum binarization threshold value is y and the average density of the image in the monitoring area is x, the above-mentioned equation (1) can be obtained from the _two points P ₂ and P ₃ .

FIG. 5 shows a procedure for setting an optimum binarization threshold value following step 1 (“ST” in the figure) in accordance with variations in illumination conditions.
1 ”) to step 23. First, in step 1 of the figure, the illuminance is set to zero, or the illuminance is set to an illuminance that enables stable measurement processing, and a model of a reference object is imaged by the TV camera 1 under this illumination condition. Then, the grayscale image is stored in the image memory 3 (step 2). In the next step 3, the operator operates the input device 16 while watching the grayscale image displayed on the video monitor 11, sets the window 23 on the grayscale image to define the above-mentioned monitoring area, and the density measuring unit 4 sets the monitoring area. The average density of the image in the monitoring area is measured (step 4).

Next, the operator causes the video monitor 11 to alternately or simultaneously display the grayscale image and the binary image,
While viewing the image, the input device 16 is operated to determine the optimum binarization threshold value. The binarized threshold value and the measured value of the average density are stored in the RAM 14 as data defining the point P ₀ or P ₂ (step 5).

Next, the illuminance is changed and set to another illuminance that allows stable measurement processing, the model of the object is imaged by the television camera 1 under this illumination condition, and the grayscale image thereof is stored in the image memory 3. It is stored (step 7). Next, in step 8, the density measuring unit 4 measures the average density of the image in the monitoring area. In the next step 9, the operator operates the input device 16 while checking the grayscale image and the binary image displayed on the video monitor 11 to determine the optimum binarization threshold value. The binarization threshold value and the measured value of the average density are RA as data defining the point P ₁ or P _3.
It is stored in M14. Next, the CPU 12 uses the RAM 14
The relational expression (1) is calculated from the data defining the points P ₀ , P ₁ or P ₂ , P ₃ stored in
Is stored in the RAM 14 (step 10).

When the measurement process is started after the relationship between the concentration and the optimum binarization threshold value is obtained in this way, step 11 becomes "YES", and stable measurement process is possible in the next step 12. Illumination conditions (for example, the illumination conditions set in step 6) are set, and the binarization circuit 6 initializes an optimum binarization threshold value under the illumination conditions manually or automatically by the CPU 12, and , And shifts to the automatic measurement process of the feature amount.

In this embodiment, whether or not the illumination condition has changed at predetermined time intervals (for example, every 5 minutes) so as to be able to cope with the change in the illumination condition caused by a change in the environment around the device or the like. It is automatically checked, and in step 13, the CPU 12 starts an internal timer.

When the image obtained by picking up the object to be recognized is input in the next step 14, C is input in the following step 15.
The PU 12 checks whether or not the timer counts for a predetermined time. If the determination is "NO", the measurement processing unit 8 measures a predetermined feature amount for the binary image of the input image stored in the image memory 7 (step 22). In the next step 23, it is determined whether or not the measurement process is completed. If the determination is “NO”, the image obtained by capturing the next recognition target object is input (step 1
4).

In this measurement process, when the timer counts a predetermined time, the determination in step 15 becomes "YES", the grayscale image is stored in the image memory 3, and then the density measuring unit 4 stores the image in the monitoring area. Is measured (step 16). If this density measurement value does not match the density measurement value stored in the RAM 14, it is determined that the illumination condition has changed, and the determination in step 17 is "YE."
S ”, and the process proceeds to step 18. C in step 18
The PU 12 applies the measured value of the density to the relational expression between the density and the optimum binarization threshold value to calculate the optimum binarization threshold value, and sets this binarization threshold value in the binarization circuit 6. (Step 19). Next, in step 20, the image of the recognition object is re-input, and the CPU 12 resets and restarts the timer (step 21).

The image input in step 20 is binarized by the binarization circuit 6 by the binarization threshold value newly set in step 19, and the process proceeds to the feature amount measurement process (step 22). . If no illumination variation is detected in step 17, go from step 17 to step 20,
The image is re-entered, and it is 2 by the binarization threshold so far.
After the value is converted and the timer is reset, the process proceeds to the measurement process (steps 21 and 22).

After the measurement procedure for all the recognition objects is completed after the same procedure is repeatedly executed, the determination in step 23 becomes "YES". The time interval for checking the variation of the illumination conditions set in the above embodiment is freely set according to the environment around the device.
Further, the method for checking the fluctuation of the lighting condition is not limited to the above method, and the lighting condition may be checked every time the measurement process is finished a certain number of times, or every time the operator inputs an external command. Then, the lighting condition may be checked.

[0032]

As described above, according to the present invention, the density of a predetermined monitoring region is measured for each image obtained by changing the illumination condition and the optimum binarization is performed under each illumination condition. The threshold value is determined, the relationship between the density and the optimum binarization threshold value is obtained in advance, and the density of the monitoring area is measured for the image obtained by imaging the object during the measurement process. By monitoring the fluctuation of the illumination condition and when detecting the fluctuation of the illumination condition from the measured value of the density, by applying the measured value of the density to the relationship between the density obtained in advance and the optimum binarization threshold value, , I tried to calculate and set the optimum binarization threshold,
There is an effect that it is possible to easily calculate and set an optimum binarization threshold value that follows changes in illumination conditions.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of the overall configuration of a visual recognition device in which the present invention is implemented.

FIG. 2 is an explanatory diagram showing a grayscale image obtained by imaging an object.

FIG. 3 is an explanatory diagram showing a relationship between a density and an optimum binarization threshold value.

FIG. 4 is an explanatory diagram showing a relationship between a density and an optimum binarization threshold value.

FIG. 5 is a flowchart showing a procedure for setting an optimum binarization threshold value.

FIG. 6 is an explanatory diagram showing a conventional method of setting a binarization threshold value.

[Explanation of symbols]

 1 TV camera 3 Image memory 4 Density measurement unit 6 Binarization circuit 12 CPU 13 ROM 14 RAM

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 7/18 W

Claims (4)

[Claims] 1. A method for setting an optimum binarization threshold value that follows changes in illumination conditions, wherein the density of a predetermined monitoring region is obtained for each image obtained by imaging under different illumination conditions. And the optimum binarization threshold under each lighting condition is determined, and the relationship between the concentration and the optimum binarization threshold is obtained in advance, and the object is imaged during the measurement process. For the image obtained by measuring the density of the monitoring region to monitor the variation of the illumination condition, when the variation of the illumination condition is detected by the measured value of the concentration, the measured value of the concentration was previously obtained. An optimum binarization threshold value setting method characterized in that the optimum binarization threshold value is calculated and set by applying the relationship between the density and the optimum binarization threshold value. 2. The densities of predetermined monitoring areas are respectively measured for an image obtained under illumination conditions when illuminance is zero and an image obtained under illumination conditions where illuminance allows stable measurement processing. The optimal binarization according to claim 1, wherein the optimal binarization threshold value under each illumination condition is determined, and the relationship between the density and the optimal binarization threshold value is obtained in advance. Threshold setting method. 3. The density of a predetermined monitoring region is measured for each image obtained under illumination conditions of different illuminance that enables stable measurement processing, and the optimum binarization threshold under each illumination condition is measured. The value is determined, and the relationship between the density and the optimum binarization threshold value is obtained in advance.
The optimum binarization threshold value setting method described in 1. 4. An apparatus for setting an optimum binarization threshold value that follows changes in illumination conditions, wherein a predetermined predetermined value is obtained for each image obtained by imaging under different illumination conditions. A storage unit that stores the relationship between the density of the monitoring area and the optimum binarization threshold value, and the density of the monitoring area of the image obtained by imaging the object during the measurement process is measured to determine the illumination condition. An optimum binarization threshold value is obtained by applying a concentration measurement unit that monitors fluctuations and a measured value of the concentration measured by the concentration measurement unit to the relationship between the concentration stored in the storage unit and the optimum binarization threshold value. Optimal 2 comprising arithmetic means for calculating a value, and binary threshold setting means for setting the optimal binary threshold calculated by the arithmetic means.
Quantization threshold setting device.