JPH11312245A - Colored foreign material detector and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a colored foreign material detector with high accuracy and simple circuit constitution which dispenses with the continuous adjustment of color balance on the side of a color video camera and detects a colored foreign material in color which is different from the base color of an object to be examined. SOLUTION: The device is provided with an A/D converted 1 which converts analog RGB image signals AR, AG and AB obtained by photographing the object to be examined into digital RGB image signals DR, DG and DB, an RGB balance adjusting means 2 for calculating the prescribed evaluated value ER, EG and EB of the signals DR, DG and DB within the prescribed area of a single image by each RGB signals and adjusting the inputting level or the reference levels RR, RG and RB of the converter 1 by each RGB, and a colored foreign material detecting means 3 detecting the colored foreign material from the photographic picture of the material to be examined based on the signals DR, DG and DB.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention detects a colored foreign substance having a color different from the base color of the inspection object from the inspection object based on an analog RGB image signal obtained by photographing the inspection object. The present invention relates to a colored foreign matter detection device and method.

[0002]

2. Description of the Related Art Conventionally, when detecting a colored foreign substance having a color different from the base color of the test object from a test object of a single color or an almost single color including an achromatic color such as white, the test object is detected. When the object to be inspected is photographed with a color video camera such as a CCD camera that has been subjected to color balance adjustment, a so-called white balance adjustment, as an adjustment target, the object to be inspected is photographed in white, and is different from the base color of the object to be inspected. Since colored foreign matter is always identified as a chromatic color, based on an analog image signal output from the color video camera,
For example, the chromatic portion is extracted from the shooting screen based on the output level of the color difference signal and the like, and the colored foreign matter is detected.

[0003]

However, when the inspection is performed while sequentially replacing the inspection object, or when the continuously moving inspection object is inspected at a fixed point, the white balance adjustment of the color video camera is performed. Is incomplete, or the color of the test object is not completely white or monochromatic, but its saturation and hue fluctuate slightly, so that the white balance may be relatively deviated. For example, there is a problem that a shadow portion on the surface of the inspection object caused by lighting or the like is recognized as the chromatic portion and malfunctions, and the malfunction also becomes more prominent due to a change in environmental conditions such as the lighting. Further, it is necessary to reduce the detection sensitivity at the expense of the detection accuracy so as not to detect such a shadow portion, or to perform the white balance adjustment of the color video camera frequently, for example, every time the inspected object changes. There was a problem that did not become.

The present invention has been made in view of the above circumstances, and has as its object to eliminate the need for continuous color balance adjustment on the color video camera side and to detect colored foreign matter with high accuracy. It is to provide a detection device and a method.

[0005]

A first feature of a colored foreign matter detecting apparatus according to the present invention for achieving this object is as follows.
An A / D converter for converting an analog RGB image signal obtained by photographing an object to be inspected into a digital RGB image signal, as described in claim 1 of the claims section, A predetermined evaluation value of the digital RGB image signal is calculated for each of the RGB signals, and the input level of the A / D converter or the input level of the A / D converter is adjusted so that the calculated evaluation value matches between the RGB signals within a predetermined error range. RGB balance adjusting means for adjusting a reference level for each of RGB, and colored foreign matter detection for detecting colored foreign matter of a color different from a base color of the object to be inspected from a photographed image of the object to be inspected based on the digital RGB image signal. Means.

[0006] In the second feature configuration, as described in claim 2 of the claims, in addition to the first feature configuration, the colored foreign matter detecting means is configured to detect the color foreign matter based on the digital RGB image signal. High-saturation pixel detecting means for sequentially calculating the saturation of each pixel to extract high-saturation pixels having a saturation equal to or higher than a predetermined reference value, and high-saturation in which the high-saturation pixels are aggregated with a predetermined density or a size or more. The point is that it comprises high-saturation pixel area detection means for detecting a pixel area. Here, the saturation is
It means the saturation of the three elements of color display, hue, lightness, and saturation. If the saturation is low, the achromatic color is mixed with the spectrum color, and the lower the saturation, the higher the degree. The spectral colors synthesized with the three primary colors of RGB include the spectral colors of actual sunlight (wavelength: about 380 nm to about 780 nm) and some complementary colors thereof.

The third characteristic configuration is the same as the third characteristic configuration, in addition to the second characteristic configuration, wherein the high-saturation pixel detecting means includes the digital RG.
The point is that a value obtained by subtracting the minimum value from the maximum value of each of the RGB signal levels of the B image signal for each pixel is calculated as the saturation for each pixel.

[0008] The fourth characteristic configuration is that, in addition to the first, second or third characteristic configuration, the RGB balance adjusting means includes:
For an RGB component having an evaluation value lower than a predetermined reference evaluation value by a predetermined amount or more, the corresponding input level is amplified by a predetermined amount or the corresponding reference level is lowered by a predetermined amount, and the predetermined reference evaluation value is reduced. For an RGB component having an evaluation value higher than a predetermined amount, the corresponding input level is attenuated by a predetermined amount, or the corresponding reference level is increased by a predetermined amount, and the evaluation value is predetermined between the RGB signals. The point is to match within the error range.

The fifth characteristic configuration is, as described in claim 5 of the claims, in addition to the fourth characteristic configuration, the predetermined reference evaluation value is calculated by the RGB balance adjusting means. That is, it is the evaluation value of a predetermined one of the evaluation values of the RGB components.

[0010] In the sixth feature configuration, as described in claim 6 of the claims, in addition to the first, second, third, or fourth feature configuration, the evaluation value is the same as that of the first, second, third, or fourth feature configuration. The point is that the signal level of the digital RGB image signal for one frame or one field in a predetermined area is an average value, a median value, or a most frequent value.

[0011] To achieve this object, a characteristic feature of the method for detecting a colored foreign matter according to the present invention is that an analog RGB image obtained by photographing an object to be inspected is described in claim 7 of the claims. An A / D conversion step of separately converting the signals into digital RGB image signals by an A / D converter, and an evaluation of calculating a predetermined evaluation value of the digital RGB image signals in a predetermined area of one screen for each of the RGB signals R and R, which adjust the input level or reference level of the A / D converter for each of the RGB signals so that the calculated evaluation values of the RGB signals match within a predetermined error range.
The GB balance adjustment step is continuously executed, and the A / D
The digital RG obtained by executing the converting step, the evaluation value calculating step, and the RGB balance adjusting step.
The present invention is characterized in that a colored foreign matter detection step of detecting a colored foreign matter having a color different from the base color of the object to be inspected from the photographed image of the object to be inspected based on the B image signal.

The operation and effect will be described below. According to the first characteristic configuration of the colored foreign matter detection device according to the present invention, on the side of the color video camera which shoots the inspection object and outputs the analog RGB image signal, the color video camera itself has poor adjustment, or Even if the color balance adjustment between the RGB components, that is, the so-called white balance is deviated due to a change in the base color of the object to be inspected, the RGB balance adjusting means adjusts the deviation of the balance adjustment of the analog RGB image signal. The input level to the A / D converter, that is, the input gain of the analog RGB image signal, or the reference level of the A / D converter is detected in order to detect the mismatch of the predetermined evaluation value and eliminate the mismatch. By adjusting each of RGB, the digital RGB image signal converted by the A / D converter always has a fixed color variation. To be maintained in the scan, the even base color of the object to be inspected is changed, the digital RGB after the variation
The base color of the inspection object reproduced by the image signal is maintained in an achromatic color or an approximately achromatic color without being changed by the fluctuation. With this, the colored foreign matter detection unit can simply identify a colored foreign matter having a color different from the base color of the inspection object as a chromatic portion regardless of the base color of the inspection object. You can. Therefore, a true colored foreign matter can be reliably detected without erroneously detecting itself as a colored foreign matter, depending on the coloring state of the inspection object itself. Also, increasing the input gain is
The same digital output value can be obtained relatively equivalent to lowering the reference level. However,
It should be noted that lowering the reference level unnecessarily lowers the noise margin during the A / D conversion processing. As described above, by automatically performing the color balance adjustment, the balance adjustment on the color video camera side needs to be performed at least once before the start, so that the subsequent adjustment is unnecessary, and the work cost for maintenance is greatly reduced. It becomes possible. Also,
Since there is no need to provide a special adjustment circuit on the apparatus side, the manufacturing cost can be reduced accordingly.
Here, the evaluation value is an index value for quantitatively evaluating an average color tone (hue, saturation, lightness) in the predetermined area, and is decomposed for each component of the RGB primary colors. I do. Therefore, by introducing such an evaluation value, the color balance adjustment can be performed easily and at high speed. In addition, by limiting the calculation of the evaluation value to a predetermined area on one screen, the predetermined area is set as an area where the object is reliably photographed, so that, for example, a background other than the object to be inspected can be used. This prevents the subject from being used for the balance adjustment. Therefore, when the object to be inspected is always photographed on the entire photographing screen, the predetermined area may be the whole photographing screen.

Incidentally, the characteristic configuration and the characteristic configuration of the method for detecting a colored foreign matter according to the present invention are essentially the same in the technical idea, so that both characteristic configurations have the same operation and effect.

According to the second characteristic configuration, even if the inspection object has color unevenness slightly different in hue or saturation from its base color, the color unevenness portion is reproduced by the digital RGB image signal. In this case, the high-saturation pixel detection unit extracts a chromatic pixel constituting the colored foreign matter as a high-saturation pixel having a saturation equal to or higher than a predetermined reference value. In addition to avoiding erroneous detection as a colored foreign substance, the high chroma pixel area detecting means removes the high chroma pixels that exist independently as a single unit or a minute set, thereby achieving the analog R
The high-saturation pixels, which may be caused by noise components superimposed on the GB image signal or the digital RGB image signal, can be effectively removed, and only true colored foreign substances can be detected with high accuracy. Further, since the high-saturation pixel detection means detects a high-saturation portion in pixel units, the digital RGB image signal can be directly processed as serial data, and further, the high-saturation pixel area detection means is sequentially detected. The high-saturation pixel area can be detected by judging the detection frequency or continuity of the high-saturation pixels to determine whether the high-saturation pixels are aggregated with a predetermined density or size or more. It is.

According to the third characteristic configuration, a color having a constant difference value obtained by subtracting the minimum value from the maximum value of the gradation values of the RGB components of the digital RGB image signal has a constant achromatic color mixing degree. By defining the same color with the same saturation, the colored foreign matter identifiable as a chromatic portion in the shooting screen can be quantitatively detected regardless of its hue. It can be done very easily. Further, from the viewpoint of digital calculation, the color difference can be detected by only the simplest calculation processing of magnitude comparison and subtraction. As a result, the colored foreign matter detecting means can be realized with a very simple circuit configuration, and the processing time thereof can be increased. Therefore, it is not necessary to use an excessively high speed and expensive one as an arithmetic element to be used, and the manufacturing cost is reduced. Can also be reduced.

Next, a description will be given of the basis for the method of calculating the saturation based on the difference value employed in the present characteristic configuration to appropriately represent the saturation. The digital RGB image signal is output from the A / D
If the converter has an N-bit resolution, 2 for each RGB signal
Can be expressed by the Nth power of A. For example, A with 8-bit resolution
If a / D converter is used, the RGB signal levels of the digital RGB image signal are 256 tones from 0 to 255. Here, the digital RGB image signal is RGB
As shown in FIG. 3, the color to be reproduced by synthesizing each component in the range of 0 to 255 is the gradation value r, g,
Let b be the coordinate value (r, g, b) of the three-dimensional rectangular coordinate system 3
It can be represented as a point in dimensional space. Also, such a reproduced color enters a cube whose coordinate values are in the range of 0 to 255,
The origin (0,0,0) is black, (255, 255, 255)
Represents white, and (255, 0, 0), (0, 25)
5,0), (0,0,255), (255,255,0),
The vertices of the cube represented by the coordinates (255, 0, 255) and (0, 255, 255) are red, green,
Represents the spectral colors of blue, yellow, magenta, and cyan. The color on the line segment L connecting the vertices in the order of red, yellow, green, cyan, blue, magenta, and red is such that at least one of the gradation values of each of the RGB components includes 0 and 255. Represents the most saturated color not including. If this is illustrated, as shown in FIG. 4, the cube becomes a hexagon consisting of the six vertices viewed from the axial direction of the achromatic axis X connecting both the vertices of black and white, and each side has the six vertices. This is the line segment L having the highest saturation connected. In addition, the center of the hexagon is achromatic, and the color becomes more vivid as it goes to the outer periphery, and the saturation becomes higher. When the maximum value of each of the RGB components of the coordinate values (R, G, B) decreases, the saturation and brightness decrease, and the color approaches black. At this time, if the gradation ratio is kept constant, the hue is kept constant; otherwise, the hue changes simultaneously. On the other hand, if the minimum value of each of the RGB components of the coordinate values (R, G, B) increases while the maximum value remains constant, the color becomes whitish and the saturation decreases. Therefore, if the minimum value is the same, the saturation decreases as the maximum value decreases, and if the maximum value is the same, the saturation decreases as the minimum value increases, so the minimum value is subtracted from the maximum value. The larger the difference value, the higher the saturation. The maximum value of the difference value is 255, which represents the color on the outer periphery of the hexagon, that is, the color on the line segment L. The minimum value is 0, and all the gradation values r, g, and b are equal to each other. A point on the achromatic axis X is shown. Here, if a set of points having a constant difference value is defined as an isochromatic surface, the isochromatic surface is composed of six surfaces except for the maximum value and the minimum value of the differential value, and each surface is individually A hexagon parallel to the achromatic axis X and similar to the hexagon shown in FIG. 4 when viewed from the axial direction of the achromatic axis X.

According to the fourth characteristic configuration, each evaluation value of each of the RGB components coincides within a predetermined error range while converging to the predetermined reference evaluation value, so that the RGB balance adjusting means can be reliably used. Therefore, the functions and effects of the first to third features can be surely exhibited. Further, by performing feedback control in which the input level or the reference level adjustment amount is changed by a constant step amount, as long as the color tone of the inspection object does not change extremely frequently, a proper number of adjustments may be made several times. Can be returned to a proper balance adjustment, so that the shooting screen can be prevented from oscillating unnecessarily, and the evaluation of each evaluation value can be performed only by a simple magnitude comparison with the predetermined reference evaluation value. Therefore, if the same effect is to be achieved by digital signal processing, the necessary multiplication or division processing becomes unnecessary, and the circuit configuration can be simplified.

According to the fifth characteristic configuration, it is unnecessary to adjust the input level or the reference level for one of the RGB components, so that the circuit configuration can be simplified.

According to the sixth characteristic configuration, the average color tone in the predetermined area can be quantitatively evaluated, and the evaluation value can be calculated by an extremely simple calculation process. Note that if the variation in the color tone of the subject related to the balance adjustment within the predetermined area is not large and the variation is random, a large difference may occur between the average value, the median value, and the most frequently occurring value. Regardless of which evaluation value is used, the same evaluation can be performed for the degree of balance adjustment as long as the same component is used for each of the RGB components.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A colored foreign matter detection according to the present invention will be described below.
An embodiment of the apparatus will be described with reference to the drawings. In FIG.
As shown, the colored foreign matter detection device according to the present invention
Shoot an object with a color video camera
Analog RGB image signal A output from the camera_R , A
_G , A_B Are input continuously for each pixel and for each pixel
To an 8-bit digital RGB image signal D_R , D_G ,
D_B A to execute the A / D conversion step of sequentially converting
/ D converter 1 and the digital RGB image signal D_R ,
D _G , D_B To maintain the required color balance,
Each reference level R of the A / D converter 1_R ,
R_G , R_B RGB balance adjusting means 2 for adjusting the
Digital RGB image signal D_R , D_G , D_B Based on
From the photographed image of the object to be inspected,
Implements a colored foreign matter detection process to detect colored foreign matter of different colors.
And a colored foreign matter detecting means 3 to be executed.

Further, the RGB balance adjusting means 2 comprises:
It comprises three average value calculation means 4 and three reference level adjustment means 5. The average value calculation means 4
The digital RGB image signals D _R , D _G , and D _B are sequentially inputted for each of RGB, and RGB signals at a central portion of a photographed image for one field, which are obtained by removing predetermined pixels from the upper, lower, left, and right outer peripheries. The average gradation value (8-bit width) of the central portion is calculated by selectively adding the gradation values (0 to 255) of
The average gradation is output for each field as the evaluation values E _R , E _G , and E _B of the digital RGB image signals D _R , D _G , and D _{B in} the one field, and an evaluation value calculation step is performed. I do. Here, the peripheral part was excluded from the target of the average gradation calculation because the object to be inspected was observed and photographed exclusively at the center of the screen, and a background of a color different from that of the object to be inspected was simultaneously photographed in the peripheral part. This is the case in which

Each of the reference level adjusting means 5 includes:
The evaluation value E_R , E_G , E_B Is entered separately, and 8
Bit width reference evaluation value E₀ See, for example, tone value 160
Input as a level and the evaluation value E_R , E_G , E_B And before
Standard evaluation value E₀ Are compared with each other, and the evaluation value E
_R , E_G , E_B Is larger than the corresponding riff
Reference level R_R , R_G , R_B Is increased by the specified increment.
The evaluation value E_R, E_G , E_B Is smaller
On the contrary, execute the RGB balance adjustment step.
You. Here, even if the step value is a minute fixed value,
Evaluation value E_R , E_G, E_B And the reference evaluation value E₀ Difference
Is a value proportional to the evaluation value E after several adjustments. _R , E_G ,
E_B And the reference evaluation value E₀ May be a value that matches
No. The reference level R_R , R_G , R_B Increase or decrease
The tone value increases or decreases in inverse proportion to the
Value E_R , E_G , E_B Is the reference evaluation value E₀ Converge on
Feedback control. Therefore, the reference
Lens level R_R , R_G , R_BAdjustment is one field lap
Time delay of one to several fields
And the digital RGB image signal D_R , D_G , D_B No ba
Lance adjustment will be performed. The reference
Level R_R , R_G , R_B The initial value of
Image signal A_R , A_G , A_B Maximum amplitude and pedestal level
Is the maximum gradation value 255.
It is set in advance as follows.

Each of the average value calculating means 4 and each of the reference level adjusting means 5 is, specifically, a microcomputer (see FIG. 1) except for an analog circuit portion for outputting each of the reference levels R _R , R _G , and R _B. (Not shown). The microcomputer also controls the timing of the colored foreign matter detecting means 3 and the like.

With the above arrangement, the digital RGB image signals D _R , D unless the fluctuation of the basic color of the object to be inspected or the fluctuation of environmental conditions such as illumination occur rapidly within a time corresponding to several fields. _G, D _B is always becomes substantially the color balance adjustment is maintained in the state they were made. As a result, the object to be inspected can always be treated as an achromatic color in image processing, and a colored foreign substance having a color different from the base color of the inspected object may be simply detected as a chromatic color. Means 3 can be realized with a simple circuit configuration as described below.

The colored foreign matter detecting means 3 sequentially calculates the saturation of each pixel based on the digital RGB image signals D _R , D _G , and D _B and obtains a high saturation of a saturation equal to or higher than a predetermined reference value C _0. It comprises high-saturation pixel detection means 6 for extracting pixels, and high-saturation pixel area detection means 7 for detecting a high-saturation pixel area in which the high-saturation pixels are aggregated at a predetermined density or higher. , FPGA (field programmable gate array) or other random logic.

As shown in FIG. 2, the high-saturation pixel detection means 6 outputs RGB input values D _R ,
D _G, subtracts the maximum value-minimum value detection unit 6a for each output picked the maximum value D _MAX and the minimum value D _MIN from D _B, the minimum value D _MIN from the maximum value D _MAX, the the subtraction value is composed of a subtraction circuit unit 6b as the chroma C of the pixel, the magnitude comparator 6c for outputting a high saturation pixel detection signal P to the reference value C ₀ and the chroma C is binarized by size comparison . That is, the high-saturation pixel detection signal P is calculated based on the saturation C
Is larger than the reference value C ₀ , the pixel is determined to be a high-saturation pixel and “1” is output. Further, as shown in FIG. 2, the maximum value / minimum value detection section 6a is composed of three 2-input magnitude comparators 6d.

The high-saturation pixel area detection means 7 sequentially inputs serial data of the high-saturation pixel detection signal P generated for each pixel, and counts "1" contained therein for each predetermined data length. If the number is equal to or more than the predetermined number, the high-saturation pixels are gathered at a predetermined density or higher, and it is determined that the pixels are a part of the colored foreign matter to be detected, and the colored foreign matter detection signal P is output.

When the colored foreign matter detection signal P is output,
Based on this signal, related post-processing such as a process for specifying the position of the colored foreign matter on the photographing screen, generating an alarm, or stopping the processing of the inspection object is performed.

Another embodiment will be described below. <1> The digital RGB image signal D_R , D_G , D_B 
The reference level is adjusted to balance each other.
Le R_R , R_G , R_B Instead of adjusting the analog
RGB image signal A_R , A_G , A_B Said A / D converter
It is also preferable to adjust the input gain to unity.
In this case, the reference level R_R, R_G , R_B 
The adjustment of the reference level R_R , R_G ,
R_B If you want to reduce the
Log RGB image signal A_R , A_G , A _B Input gain
Quantitatively amplify the reference level R_R , R_G , R_B 
When the input gain is increased by a predetermined step value,
What is necessary is just to attenuate by a predetermined amount. However, the evaluation value E_R , E
_G , E_B And the reference evaluation value E₀ The amount equivalent to the difference of
Because there is a risk of oscillation if you try to adjust it,
Reference level R_R , R_G , R_B Adjustment amount
It is preferable to carry out by dividing.

<2> In the above embodiment, the reference evaluation value E ₀ is a fixed value of the gradation value 160, but may be another gradation value. It does not matter. For example, as the reference evaluation value E _0, the evaluation value E _R, E _G, may be used any one of E _B. For example, if the evaluation value E _G was the reference evaluation value E ₀ is the reference level R _R, R _G, the adjustment of R _B by fixing the reference level R _G, other 2
One of the reference level R _R, it is also preferred embodiment to be performed on R _B. By doing so, the reference level adjustment or the input gain adjustment can be omitted for each of the RGB components, and the circuit configuration can be simplified.

<3> The RGB balance adjusting means 2
The digital RGB
Although the average gradation of the predetermined area of the image signals D _R , D _G , and D _B is calculated and used as the evaluation values E _R , E _G , and E _B , the predetermined area is calculated instead of calculating the average gradation. The median value or the most frequent value of the distribution of the gradation values within may be used as the evaluation value.

<4> The high-saturation pixel area detection means 7 sequentially inputs serial data of the high-saturation pixel detection signal P generated for each pixel, and detects that "1" is continuously generated for a predetermined data length. Thus, the continuous data may be determined to be a high-saturation pixel appearing on one scanning line of the colored foreign matter, and the colored foreign matter detection signal P may be output.

<5> The RGB balance adjusting means 2
The balance adjustment was performed every field
Changes the processing range from one field to one frame,
It may be performed for each frame. Also, be sure
Also, for each field or frame, the balance
Fees that do not perform the appropriate processing without performing the adjustment processing
A frame or a frame may be provided. Also, multiple units
Sets of analog RGB images from a color video camera
Image signal A _R , A_G , A_B The one colored foreign matter detection device
It is also preferable to perform a time division process. This place
In this case, the balance for each field or frame
Perform the adjustment process in the field or frame for the number of cameras.
It is also preferable to perform the cyclic processing at a cycle of the system. With this configuration
If the object to be processed is photographed from another direction
You can.

<6> Each of the A / D conversion step, the evaluation value calculation step, the RGB balance adjustment step, and the colored foreign matter detection step is performed by a configuration other than the colored foreign matter detection device according to the present invention. Each of the above steps may be executed by an independent device. Alternatively, a part of the steps, for example, the A / D conversion step, the evaluation value calculating step, and the R
The color video camera performs a GB balance adjustment step on the color video camera side.
An embodiment that outputs the B image signals D _R , D _G , and D _B may be used.

[0035]

As described above, according to the present invention,
A colored foreign matter detection device that does not require continuous color balance adjustment on the color video camera side and can detect colored foreign matter of a color different from the base color of the inspection object with high accuracy and a simple circuit configuration Could be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a colored foreign matter detection device according to the present invention.

FIG. 2 is a main block diagram showing an embodiment of a colored foreign matter detection device according to the present invention;

FIG. 3 is an explanatory diagram for explaining color display using digital RGB image signals.

FIG. 4 is an explanatory diagram for explaining saturation in color display using digital RGB image signals.

[Explanation of symbols]

Reference Signs List 1 A / D converter 2 RGB balance adjusting means 3 Colored foreign matter detecting means 4 Average value calculating means 5 Reference level adjusting means 6 High chroma pixel detecting means 6a Maximum / minimum value detecting unit 6b Subtraction circuit unit 6c Large / small comparator (1-bit output) 6d Large / Small comparator (8-bit output) 7 High-saturation pixel area detection means A _R , A _G , A _B analog RGB image signal D _R , D _G , D _B digital RGB image signal R _R , R _G , R _B reference level E _R , E _G , E _B evaluation value E ₀ reference evaluation value P High chroma pixel detection signal C Saturation C ₀ reference value D _MAX maximum value D _MIN minimum value

Claims (7)

[Claims] 1. An analog R obtained by photographing an object to be inspected.
A for converting a GB image signal into a digital RGB image signal
A / D converter, and a predetermined evaluation value of the digital RGB image signal in a predetermined area of one screen is calculated for each of the RGB signals.
RGB balance adjusting means for adjusting the input level or reference level of the A / D converter for each of RGB so that the calculated evaluation value matches within a predetermined error range between the signals, and A colored foreign matter detecting device for detecting a colored foreign matter having a color different from a base color of the inspected object from the photographed image of the inspected object based on the detected image. 2. A high-saturation pixel detection means for sequentially calculating the saturation of each pixel based on the digital RGB image signal and extracting a high-saturation pixel having a saturation equal to or higher than a predetermined reference value. 2. The colored foreign matter detection device according to claim 1, further comprising a high-saturation pixel area detecting means for detecting a high-saturation pixel area in which the high-saturation pixels are aggregated with a predetermined density or a size or more. 3. The high-saturation pixel detection means calculates, as the saturation of each pixel, a value obtained by subtracting a minimum value from a maximum value of each of RGB signal levels of the digital RGB image signal for each pixel. The colored foreign matter detection device according to claim 2, wherein 4. The method according to claim 1, wherein the RGB balance adjusting means amplifies the corresponding input level by a predetermined amount with respect to an RGB component having an evaluation value lower than a predetermined reference evaluation value by a predetermined amount or more.
Alternatively, the corresponding reference level is lowered by a predetermined amount, and for the RGB component having an evaluation value higher than the predetermined reference evaluation value by a predetermined amount or more, the corresponding input level is attenuated by a predetermined amount, or 4. The colored foreign matter detection device according to claim 1, wherein the reference level is increased by a predetermined amount, and the evaluation values are made to match within a predetermined error range between the RGB signals. 5. The colored foreign matter detection device according to claim 4, wherein the predetermined reference evaluation value is an evaluation value of a predetermined one of the evaluation values of each of the RGB components calculated by the RGB balance adjusting means. 6. The method according to claim 1, wherein the evaluation value is an average value, a median value, or a most frequent value of signal levels of the digital RGB image signal for one frame or one field in the predetermined area. 6. The colored foreign matter detection device according to 3, 4, or 5. 7. An analog R obtained by photographing an object to be inspected.
A / D converter converts each of RGB image signals to digital RGB
An A / D conversion step of converting the image signal into an image signal; an evaluation value calculation step of calculating a predetermined evaluation value of the digital RGB image signal in a predetermined area of one screen for each of the RGB signals; And an RGB balance adjusting step of adjusting an input level or a reference level of the A / D converter for each of RGB so that the evaluated values match within a predetermined error range. Based on the digital RGB image signal obtained by performing the evaluation value calculating step and the RGB balance adjusting step, a color different from the base color of the test object from the photographed image of the test object. A colored foreign matter detection method, comprising performing a colored foreign matter detection step of detecting a colored foreign matter.