JPH05314252A - Color converting device and object inspecting device - Google Patents

Abstract

PURPOSE:To properly convert red, green and blue signals to hue, saturation and lightness signals without being affected by any environmental change by converting the red, green and blue signals according to a correctable conversion function and outputting those signals. CONSTITUTION:The color signals in the three colors of red, green and blue inputted from a camera 3 receive the compensation of half tone by non-linear conversion tables 12R, 12G and 12B and are respectively converted to linear characteristics. The tables 12R-12B are look-up tables to be rewritten by a CPU 16. The values of the tables 12R-12B are set from a ratio between the values of the red, green and blue signals obtained by photographing a gray sample and the original red, green and blue values of the gray sample. The red, green and blue signals outputted from the tables 12R-12B are converted to the signals of hue H, saturation S and lightness I by an RGB-H conversion part 13H, RGB-S conversion part 13S and RGB-I conversion part 13I.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color conversion device for converting red, green and blue signals into hue, saturation and brightness signals and an object inspection device using the same.

[0002]

2. Description of the Related Art Conventionally, based on a predetermined conversion function,
The hue, saturation, and brightness signals that are uniquely determined from the red, green, and blue signals are obtained.

[0003]

However, even if the same color sample is used, the characteristics of the input device such as a camera that converts it into a video signal, the ambient brightness when the image is captured, and the color temperature of the illumination. Changes the values of red, green, and blue,
The values of hue, saturation, and brightness also change.

In order to correct it, white balance adjustment of a camera is generally performed. However, they only give linear compensation and are done manually.

As a result, changes in the input image must be absorbed in image processing, and the changes may not be linear, which adversely affects accuracy and processing time.

The present invention has been made in view of such circumstances, and makes it possible to correctly convert red, green, and blue signals into hue, saturation, and lightness signals without being affected by environmental changes. With the goal.

[0007]

According to a first aspect of the present invention, there is provided a color conversion device for converting red, green and blue signals according to a correctable conversion function and outputting the converted signals (for example, the linear conversion section 10R of the embodiment). 10G and 10B, and non-linear conversion tables 12R, 12G and 12B), and second conversion means (for example, the second conversion means for converting the red, green and blue signals output from the first conversion means into hue, saturation and lightness signals). The RGB-H conversion unit 13H, the RGB-S conversion unit 13S, and the RGB-I conversion unit 13I) of the embodiment are provided.

According to another aspect of the color conversion device of the present invention, the first conversion means receives the red, green and blue signals and outputs the red, green and blue conversion table values stored corresponding to these signals. Possible conversion table (for example, the non-linear conversion tables 12R, 12G and 12B of the embodiment), and the values of this conversion table are the values of the red, green and blue signals obtained by photographing the gray sample and the original value of the gray sample. It is characterized in that it is set from the ratio with the red, green and blue values of.

An object inspection apparatus according to a third aspect of the present invention is a first conversion means (for example, the linear conversion of the embodiment) which converts and outputs red, green and blue signals obtained by photographing an object according to a correctable conversion function. Parts 10R, 10G and 1
0B, and non-linear conversion tables 12R, 12G and 12B), and second conversion means (for example, in the embodiment) for converting the red, green and blue signals output from the first conversion means into hue, saturation and lightness signals. RGB-H converter 13
H, RGB-S converter 13S and RGB-I converter 1
3I) and determining whether the color of the object is appropriate based on the output of the second conversion means.

In the object inspection apparatus according to the present invention, the first conversion means receives the red, green and blue signals and outputs the red, green and blue conversion table values stored corresponding to these signals. A rewritable conversion table (for example, the non-linear conversion tables 12R, 12G and 12B of the embodiment), and the values of the conversion table are the values of red, green and blue signals obtained by photographing a gray sample and the gray sample It is characterized by being set from the ratio with the original red, green and blue values of.

[0011]

In the color conversion device having the structure of claim 1, red,
The green and blue signals are converted and output according to the correctable conversion function. Therefore, correct color conversion can be performed by correcting the conversion function according to the change in the environment.

According to another aspect of the color conversion device of the present invention,
The value of the rewritable conversion table is set from the ratio of the red, green and blue signal values obtained by photographing the gray sample and the original red, green and blue values of the gray sample.
Therefore, the conversion function can be corrected non-linearly according to the change of environment, and the color conversion can be performed more accurately.

In the object inspection apparatus of the third aspect, the red, green and blue signals are converted and output according to the correctable conversion function. Therefore, by correcting the conversion function according to the change in the environment, it is possible to correctly determine whether or not the color of the object is appropriate.

In the object inspection apparatus having the structure of the fourth aspect, the values of the rewritable conversion table are the values of red, green and blue signals obtained by photographing the gray sample and the original red, green and blue of the gray sample. It is set from the ratio with the value of. Therefore, it is possible to correct the conversion function non-linearly according to changes in the environment, and it is possible to more accurately determine whether or not the color of the object is appropriate.

[0015]

FIG. 1 shows an embodiment of the object inspection apparatus of the present invention. This example is an apparatus that inspects whether or not the color of the inspection target object on the belt conveyor is correct. The inspection object 1 is placed on the conveyor 2 and inspected one after another. The video camera 3 captures an image of the object 1, and the image processing device 4 determines the color of the object 1 based on the output signal of the camera 3 and asks the operator whether the monitor TV 5 is good or defective. Inform. In this embodiment, lighting 6 such as a fluorescent lamp is used.
By changing the color conversion characteristic of the image processing device 4 using the known color sample 7 with respect to the fluctuation of the brightness of
The effect can be removed.

FIG. 2 shows an example of the configuration of the image processing device 4 shown in FIG. Input signals from the camera 3 are three color signals of red, green and blue. These are input to the linear conversion units 10R, 10G, and 10B, respectively. The linear conversion units 10R, 10G, and 10B use the white color sample 7
Are used to adjust the red, green, and blue colors to the same signal level. The linear conversion units 10R, 10G, and 10B can be configured by, for example, variable gain amplifiers, and are adjusted by a signal from the output port 17 of the CPU 16.

The output signals of the linear converters 10R, 10G and 10B are input to the A / D converters 11R, 11G and 11B, respectively, and converted into digital signals.
In this example, 8-bit quantization, that is, conversion into a digital signal from 0 to 255 is performed. A / D converter 11R, 1
The output signals of 1G and 11B are compensated for intermediate colors by the non-linear conversion tables 12R, 12G and 12B, respectively, and converted into linear characteristics. The non-linear conversion tables 12R, 12G and 12B are lookup tables that can be rewritten by the CPU 16.
For rewriting, the CPU 16 and the non-linear conversion tables 12R, 12G and 12B are connected by an address and data bus.

The red, green and blue signals output from the nonlinear conversion tables 12R, 12G and 12B are RGB-
H conversion unit 13H, RGB-S conversion unit 13S and RGB
The -I conversion unit 13I causes the hue, the saturation, and the brightness (H,
S, I) signals are converted. A lookup table is used for these conversions, and RGB-Lab, RGB-
Conversion of Munsell etc. is performed. By controlling the output port 17 of the CPU 16, the RGB-H conversion units 13H, RGB
-S conversion 13S and RGB-I conversion unit 13I are not used, and non-linear conversion tables 12R, 12G, and 12B are used.
It is possible to bypass the red, green and blue signals output by the.

The hue, saturation and lightness signals (data) output from the RGB-H converter 13H, RGB-S converter 13S and RGB-I converter 13I are stored in frame memories 14H, 14S and 14H, respectively. It The CPU 16 obtains the hue, saturation, and brightness data of the inspection object 2 from the frame memories 14H, 14S, and 14H, compares them with the data stored in the program / data memory 18 in advance, and compares the result with the frame memory 1
9H, 19S and 19I, and the HSI-R converter 20R, HSI-G converter 20G and HSI-B converter 20B restore the red, green and blue signals to the D / A converters 21R, 21G and 21B. To monitor TV via.

FIG. 3 shows the overall flow of processing by the CPU 16. The processing is large, and the color conversion circuit is adjusted and the color of the object 1 to be inspected is stored (step S1).
And the operation (step S2) of actually injecting the target object 1 onto the belt conveyor 2 to inspect.

FIG. 4 shows the flow of color adjustment processing in step S1 shown in FIG. After the initialization (step S11), the white level is adjusted to effectively use the dynamic range (step S12), and the intermediate level is adjusted to remove the non-linear factor (step S13).
The color of the inspection object 1 is stored (step S14).

FIG. 5 shows a specific example of the initialization shown in step S11 of FIG. First, the linear conversion units 10R, 10
The gains of G and 10B are set to 1 (step S2
1). Then, the outputs of the nonlinear conversion tables 12R, 12G, and 12B are made equal to the inputs (step S2).
2). By this initialization, the input signal is subjected to color conversion without any processing.

FIG. 6 shows a concrete example of the white level adjustment shown in step S12 of FIG. First, the RGB-H conversion units 13H and R are controlled by the output port 17 of the CPU 16.
The signal is bypassed without using the GB-S converter 13S and the RGB-I converter 13I (step S3
1). Then, a white sample of Munsell standard color chart is photographed by the camera 3 (step S32), and adjustment is performed so that the digital signals after A / D conversion become the highest level for each of red, green and blue. In the case of 8-bit quantization, 255 is the highest level, so in order to judge whether or not it exceeds 254, it is adjusted to 254, which is 1 smaller than that.

That is, the red signal of the image data of the white sample is checked (step S33), and if it is smaller than 254, the gain of the red channel linear conversion unit 10R is increased (step S34), and if it is 255, the gain is increased. Lower (step S35). Similarly, the level of the green signal is checked (step S36), and if it is smaller than 254, the gain of the green channel linear conversion unit 10G is increased (step S3).
7) If 255, decrease gain (step S3)
8). The same applies to the green signal (step S39,
S40 and S41).

FIG. 7 shows a specific example of the intermediate level adjustment shown in step S13 of FIG. In this process, the non-linearity of the camera 3 and the signal processing system is removed. even here,
Following the white level adjustment, output port 17 of CPU 16
The control bypasses the signal without using the RGB-H conversion unit 13H, the RGB-S conversion unit 13S, and the RGB-I conversion unit 13I. First, several kinds of Munsell standard color samples are prepared. The gray value and the number thereof are arbitrary, but the higher the value, the higher the accuracy.

First, one of the gray samples is taken by the camera 3
The image is taken with (step S51). Then, the levels of the red signal, the green signal, and the blue signal at that time are stored (steps S52, S53, and S54). This process is repeated for the number of samples (step S55).

Next, the ratio with the value to be originally stored (the original value of the gray sample) is calculated from the stored value (step S56). This is the conversion table 12R,
The values are 12B and 12G. Since the calculated value has only the number of gray samples, the rest needs to be interpolated. In this example, the Bezier curve is approximated (step S5).
7). Then, the approximated values are converted into conversion tables 12R, 12
Write to B and 12G (step S58).

8 to 10 show step S56 of FIG.
The process from S58 to S58 will be described in detail. Figure 8
Shows the initial state of the conversion tables 12R, 12B and 12G. The output values of the conversion tables 12R, 12B and 12G are equal to the input values. FIG. 9 shows a plurality of gray samples taken and converted into conversion tables 12R, 12B and 12G.
The process of obtaining the value of is shown. In FIG. 9, black circles are values obtained by photographing a gray sample with a camera. ×
Indicates a value that should be originally obtained, that is, a target value. That is, the black circles indicate the coordinates (input, shooting value),
X indicates coordinates (input, target value). Therefore, in order to obtain (input, target value), the $ target value sub photographing value $ may be output for that input (triangle mark). Do some of this. FIG. 10 shows the result of interpolating these. This is the non-linear conversion table 12R, 12B
And write to 12G.

FIG. 11 shows a specific example of the color storage processing of the object to be inspected in step S14 of FIG. In this process, the inspection target object 1 is actually photographed by the camera 3 and its color is stored. First, the bypass of the RGB-H conversion unit 13H, the RGB-S conversion unit 13S, and the RGB-I conversion unit 13I, which has been performed so far, is stopped (step S61), and the inspection target object 1 is photographed by the camera 3 (step S62). , The color is stored as H, S, I values (step S63).

FIG. 12 shows a specific example of the operation process shown in step S2 of FIG. In this process, the inspection object 1 is inspected one by one because it is placed on the belt conveyor 2 and is conveyed one after another. Here, it is necessary to check whether the conditions such as the illumination 6 have changed, and if they have changed, it is necessary to correct them, so that not only the inspection target object 1 but also the white sample 7 is flowed at the same time.

First, it is judged whether the object photographed by the camera 3 is the inspection object 1 or the color sample 7 (step S71). If it is the inspection object 1, the HSI value is compared with the value stored in the inspection object color storage process (step S7).
2 and S73), if they are the same, a non-defective product (step S7
4) If they are different, it is determined to be a defective product (step S75).

When the color sample 7 is not the object 1 to be inspected, the RGB values of the sample are compared to see if they are 254 (step S77). If they are different, the linear conversion units 10R, 10G, and 10B are adjusted in the same manner as the white level adjustment (step S78).

Although an example using a white sample is shown here, a gray sample may be used. It is also possible to have a configuration in which the sample can be checked at the same time as the inspection target object 1.

According to the above-described embodiment of the present invention, it is not necessary to recognize the difference and absorb the change in the image processing algorithm after the correction by the color sample. In addition, the correction can be performed non-linearly. That is, according to the above embodiment,
You get the following benefits: 1) The processing time of image processing can be shortened and the number of processing items can be reduced. 2) The accuracy of color conversion can be improved by normalizing the values of red, green, and blue before color conversion. 3) Non-linear correction can be performed, and flexibility can be dealt with. 4) It can be automatically adjusted by computer control from the image processing device.

[0035]

According to the color conversion device of the first aspect, the red, green and blue signals are converted and output in accordance with the correctable conversion function, so that the conversion function is corrected according to the change of the environment. Thus, color conversion can be performed correctly.

According to the color conversion device of the second aspect, the values of the rewritable conversion table are set to the values of the red, green and blue signals obtained by photographing the gray sample and the original red, green and blue of the gray sample. Since the setting is made from the ratio with the value, the conversion function can be corrected non-linearly according to the change of the environment, and more accurate color conversion can be performed.

According to the object inspection apparatus of the third aspect, the red, green and blue signals are converted and output according to the correctable conversion function. Therefore, by correcting the conversion function according to the change of the environment, It is possible to correctly judge whether or not the color of the object is appropriate.

According to the object inspection apparatus of the fourth aspect, the values of the rewritable conversion table are the values of the red, green and blue signals obtained by photographing the gray sample and the original red, green and blue of the gray sample. Since the setting is made based on the ratio with the value, the conversion function can be non-linearly corrected according to the change of the environment, and it can be more accurately determined whether the color of the object is appropriate.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram showing a configuration of an embodiment of an object inspection apparatus of the present invention.

2 is a block diagram showing a configuration example of an image processing device 4 of FIG. 1, that is, a configuration of an embodiment of a color conversion device of the present invention.

FIG. 3 is a flowchart showing the overall processing of the image processing apparatus of FIG.

4 is a flowchart showing an example of color adjustment processing of the image processing apparatus of FIG.

5 is a flowchart showing an example of initialization processing of the image processing apparatus of FIG.

FIG. 6 is a flowchart showing an example of white level adjustment processing of the image processing apparatus of FIG.

7 is a flowchart showing an example of an intermediate level adjustment process of the image processing apparatus of FIG.

FIG. 8 shows nonlinear conversion tables 12R, 12G and 12
It is a figure which shows the initial state of B.

FIG. 9 is a diagram showing a process of taking a plurality of gray samples and obtaining a conversion table value.

FIG. 10 is a diagram showing conversion table values interpolated by a Bezier curve.

11 is a flowchart showing an example of color storage processing of a target object of the image processing apparatus of FIG.

12 is a flowchart showing an example of the operation of the image processing apparatus of FIG. 2, that is, an example of an object color inspection process.

[Explanation of symbols]

 1 Object to be inspected 3 Video camera 4 Image processing device 5 Monitor television 7 Color sample 10R, 10G, 10B Linear conversion unit 12R, 12G, 12B Non-linear conversion table 13H RGB-H conversion unit 13S RGB-S conversion unit 13I RGB-I conversion Part 16 CPU

Claims (4)

[Claims] 1. A color conversion device for converting red, green and blue signals into hue, saturation and lightness signals, and first conversion means for converting and outputting the red, green and blue signals according to a correctable conversion function, A color conversion device comprising: second conversion means for converting the red, green, and blue signals output from the first conversion means into hue, saturation, and lightness signals. 2. The first conversion means receives the red, green and blue signals and stores the red signals corresponding thereto,
It has a rewritable conversion table that outputs green and blue conversion table values, and the values of this conversion table are the red, green and blue signal values obtained by photographing a gray sample, and the original red and green values of the gray sample. The color conversion device according to claim 1, wherein the color conversion device is set based on a ratio between the color conversion value and the value of blue. 3. An object inspection apparatus for inspecting whether or not the color of an object is proper, wherein the red, green and blue signals obtained by photographing the object are converted and output according to a correctable conversion function.
A conversion means; and a second conversion means for converting the red, green, and blue signals output from the first conversion means into a hue, saturation, and lightness signal, and the object based on the output of the second conversion means. An object inspection device characterized by determining whether or not the color of the object is appropriate. 4. The first conversion means receives the red, green and blue signals and stores the red signals corresponding to these signals,
It has a rewritable conversion table that outputs green and blue conversion table values, and the values of this conversion table are the red, green and blue signal values obtained by photographing a gray sample, and the original red and green values of the gray sample. 4. The object inspection apparatus according to claim 3, wherein the object inspection apparatus is set based on a ratio between the value of the color and the value of blue.