JP2807031B2 - Color image signal evaluation method and color image signal evaluation apparatus using this evaluation method - Google Patents

Description

The present invention relates to a method for evaluating a color image signal suitable for testing a color image element made of, for example, a semiconductor, and a method for evaluating the color image signal. The present invention relates to a color image signal evaluation device.

"Prior art" An image sensor made of a semiconductor is made by an integrated circuit technique, but various defects occur due to defects in the manufacturing process. In particular, when a color filter and an image sensor are combined to assemble a color image device, due to a defect of the image sensor, jading shown in FIG. 6, oblique stripes shown in FIGS. 7 and 8, stripes shown in FIG. Color unevenness occurs.

In other words, the screen shown in FIG. 6 should be white and a wide area of the screen should be white (this is indicated by reference numeral 1).
Refers to color unevenness that is colored over

The oblique stripes shown in FIG. 7 and FIG. 8 indicate a phenomenon in which obliquely colored stripes 2 and 3 occur on a screen that should be white.

The stripes shown in FIG. 9 indicate a phenomenon in which, for example, stripes 4 occur in the vertical direction.

Conventionally, a color image signal is projected on a color cathode ray tube, and this image is visually observed to detect color unevenness.

As described above, in the related art, the color unevenness generated in the color image pickup device is artificially determined, so that the efficiency is low. Particularly, in a mass production factory, a large number of inspectors must be arranged, which is disadvantageous in that it goes against labor saving and does not lead to cost reduction.

In addition, when the color unevenness clearly appears among the defects, it can be detected without individual difference, but when the color unevenness appears only thinly, it may be overlooked due to the individual difference of the inspector.

"Prior art" From this point, the applicant of the present invention (Japanese Patent Application No. 1-233669).
No .: a color image signal evaluation method and a color image signal evaluation apparatus using this evaluation method).

According to the invention proposed earlier, it is possible to mechanically quantify the degree of color unevenness, and it is possible to provide a test apparatus that automatically determines the presence or absence of color unevenness.

The outline of the previously proposed color image signal evaluation method and a color image signal evaluation apparatus using this color image signal evaluation method will be described with reference to FIGS. 4 and 5. FIG.

In FIG. 4, reference numeral 10 denotes a color image data source. This color image data source 10 has, for example, three memories 10R, 10G,
It can be configured by 10B. These three memories 1
For example, 0R, 10G, and 10B separate a color image signal captured by a color image sensor into red, green, and blue single-color signals, and convert the separated single-color signal into an AD-converted red image data R, green image data G,
The blue image data B is stored.

The color image data R, G, B read from the color image data source 10 is given to a conversion means 20, and the conversion means 20 converts the image data into hue image data H, lightness image data L, and saturation image data S. . As the conversion means 20, for example, HLS conversion means or HSV conversion means can be used.

The hue image data H, the brightness image data L, and the saturation image data S obtained by the conversion by the conversion means 20 are stored in the hue image memory 30A, the brightness image memory 30B, and the saturation image memory 30C, respectively.

“Hue” is obtained as a numerical value in the hue image memory 30A, “brightness” is obtained as a numerical value in the brightness image memory 30B, and “color vividness” is obtained as a numerical value in the chroma image memory 30C.

In the invention proposed earlier, each of these image memories 30
Focusing particularly on the saturation image data S among the image data H, L, S stored in A, 30B, 30C, the saturation image data S is subdivided into local blocks KB composed of m × n pixels, Each local block
Perform processing to find the average saturation in KB.

The local blocks KB perform this process over the entire screen without overlapping each other. The number of pixels m × n is given as about 10 × 10 as an initial value.

When this process is performed over the entire screen, the maximum value of the average saturation of each local block KB is extracted by the calculation unit 40, and the calculation unit 40
The maximum value of the average value extracted at 40 is sent to the adding unit 50 and stored.

The calculation unit 40 calculates the number of pixels m × n of the local block KB as m = m +
Change to k ₁ , n = n + k ₂ and again local block KB of the whole screen
Find the average value of The calculation unit 40 extracts the maximum value of the average value of the average saturation of each local block KB, and sends it to the addition unit 50.

Each time the number of pixels of the local block KB is changed, the average saturation of the local block KB on the entire screen is obtained, the maximum value is extracted, and the addition is performed by the adding unit 50.

The process is repeated until the number of the local blocks KB becomes, for example, four in the screen. The numerical value obtained by the adding unit 50 at the end of the processing is the sum of the maximum values of the average saturation. It can be determined that the color image signal has less color unevenness as the sum value is smaller.

FIG. 5 shows a graph in which the maximum value of the local block KB at each local block size is plotted based on the actually measured data. If the curve A, B, and C show characteristics close to a monotonically decreasing function as shown in the graph, it is a good product and the curves D, E, and F
If the function does not become a monotonically decreasing function as in the above, it is a defective product.

The value obtained by the adding unit 50 indicates the height of these curves A to F, and the lower the position of the curve, the smaller the sum value.

"Problems to be Solved by the Invention" In the previously proposed invention, the presence or absence of color unevenness was determined by focusing only on the saturation value. Differences in color reproducibility occur due to differences in the characteristics of color filters provided on the front surface of the image sensor (differences in the types of filters), whereby certain colors are emphasized due to differences in the types of color image sensors and the like.
There is a phenomenon that it is lower than other colors.

When the presence or absence of color unevenness is determined only by the saturation value, even if the color saturation value is high due to the presence of color unevenness of a color that is hardly perceived by the human eye, it is determined that “color unevenness is present”. There are drawbacks.

The present invention has been made to solve this drawback.

[Means for Solving the Problems] According to the present invention, when obtaining the average value of the saturation data in the local block of the saturation image data in the saturation image memory, the saturation of each pixel constituting each of the local blocks is obtained. Degree value,
A color image signal evaluation method that corrects with a color correction value calculated based on a hue value at the same pixel position, modifies a calculated value of average saturation depending on a color, and allows a pass / fail judgment criterion to be arbitrarily set for each color. And an apparatus using this method.

"Embodiment" A method for evaluating a color image signal according to the first invention of this application will be described with reference to FIG.

In FIG. 1, parts corresponding to those in FIG. 4 are denoted by the same reference numerals.

In the present invention, when calculating the average saturation in the local block KB in the saturation image memory 30C, the local block KB
The hue value reading means 31 reads out the hue value at the same pixel position as the pixel constituting the pixel, and calculates the color correction value by the color correction calculating means 32 based on the hue value. Is corrected by the correction means 33, the corrected saturation value is given to the average saturation calculation means 34, and the average saturation calculation means 34 calculates the average saturation of each local block KB.

The calculation of the color correction value in the color correction value calculation means 32 is, for example, F: color correction value A, B: coefficient h: hue value

As an example, N = 2, A = 2.0, A ₁ = 1.0, A ₂ = 0.5, B ₁
FIG. 2 shows the characteristics of the color correction values F when B = 10 and B ₂ = 300. By appropriately setting the values of N, A, and B, the correction characteristics shown in FIG. It can be set so that

At the same time as reading the saturation values of the pixels constituting each local block KB in the saturation image memory 30C, the hue value h of the hue image memory 30A corresponding to each pixel is read, and the color correction value F is obtained from the hue value h. Then, the color correction value F is multiplied by the saturation value, the multiplied result is given to the average saturation calculation means 34, and the average saturation of each local block KB is calculated using the corrected saturation value.

When the average saturation of the local block KB having the number of pixels m × n is calculated over the entire screen, the maximum value among them is selected by the maximum value selection means 35, and the maximum value of the selected average saturation is added. Input to the unit 50.

At the same time, the number of pixels m × n of the local block KB is calculated as m + k ₁ ,
Change to n = n + k _2, the average saturation of each local block KB of the number of pixels calculated with color correction, added to the previous value sent to the adder unit 50 and selects the highest value among them. This operation is repeated until the local block KB has a size of about 1/4 of the screen.

At the time when the size of the local block KB becomes about 1/4 of the screen, the operation of calculating the average saturation of the local block KB is finished. The sum of the maximum values of the block KB is compared with a reference value, and if the sum of the maximum values exceeds the reference value, it is determined to be defective, and if the sum is less than the reference value, it is determined to be good.

As described above, according to the first invention of this application, when calculating the average saturation, the saturation value of each pixel is corrected by the color correction value obtained from the hue value of the pixel position,
It is possible to make a determination as to whether or not the quality matches the sensitivity characteristic for the color of the eyes or the characteristic of a color filter provided on the front surface of the image sensor.

 FIG. 3 shows an embodiment of the second invention of this application.

In this embodiment, the hue value reading means 31, the hue value calculating means 32, the correcting means 33, the average saturation calculating means 34 shown in FIG.
The maximum value storage means 35, the determination means 51, and the control means for controlling the number of pixels of the local block KB are all changed to a computer.
60 shows the case where it is configured.

As is well known, the computer 60 includes a central processing unit 61.
And a ROM 62 containing a program for operating the central processing unit 11 in a predetermined order, a RAM 63 for taking in data and the like, an input port 64, and an output port 65.

Hue image memory 30A and saturation image memory 3 for input port 64
Provide 0C read output. The computer 60 divides the read saturation image data into local blocks KB, and reads from the hue image memory 30A a hue value at the same pixel position as a pixel constituting each local block KB.

The hue data read from the hue image memory 30A is stored in RAM63.
The central processing unit 61 performs arithmetic processing according to the first expression stored in the ROM 62 in advance, and calculates a hue correction value F corresponding to the hue value.

After calculating the color correction value F, the central processing unit 61 multiplies the color correction value F by the saturation data of the same pixel, corrects the saturation value, and uses the corrected saturation value for each local block. The average saturation is determined for each of the colors, and the maximum value among the average saturations is selected and stored in the maximum value storage area provided in the RAM 63.

The central processing unit 61 calculates the number of pixels m × n constituting the local block KB that subdivides the saturation image memory 30C into m + k1,
Conversion to n + k ₂ and the calculation of the average saturation is repeated until the local block KB becomes about / 4 of the screen.

The maximum value of the average saturation for each size stored in the RAM 63 maximum value storage area is cumulatively added, and the cumulatively added value is compared with a reference value.If the added value exceeds the reference value, the display 70 is displayed. The defect is displayed, and if it is equal to or less than the reference value, the display 70 indicates that it is good. Further, the pass / fail indication signal can be used for selecting the device under test.

[Effect of the Invention] As described above, according to the present invention, when calculating the average saturation of a local block of each size, the saturation value of each pixel is corrected with the color correction value calculated based on the hue at the same position as that pixel. Therefore, the detection sensitivity for determining a defect can be made different depending on the color that occurs as color unevenness.

In particular, by appropriately setting the value of n shown in the first equation, the color correction value F can be set to an arbitrary value for each color, for example, as shown by the curve in FIG. The criterion of pass / fail can be set for each color according to the color characteristics of the color signal.

Therefore, for example, even if the average saturation shows a high value, the color unevenness of a color that is not so noticeable to human eyes is judged to be good, and the color unevenness appears only faintly. In the case where the image appears to be emphasized to the eyes, it can be determined to be defective, and an improvement in yield can be expected in a test of a color image sensor or the like.

In the above description, an example in which the color correction value is calculated for each pixel has been described.
The same local block K in the hue image memory 30A corresponding to
B may be used, and the color correction value F may be obtained from the average saturation of the local block KB using the first equation.

Further, a curve for determining the color correction value F may be stored in a memory in advance, and a curve having various characteristics may be called up as necessary to perform hue value-color correction value conversion.

[Brief description of the drawings]

FIG. 1 is a functional explanatory diagram for explaining a color imaging signal evaluation method proposed in the first invention of the present application, and FIG. 2 is a diagram for explaining an example of a characteristic of a color correction value used in the first invention. FIG. 3 is a block diagram showing an embodiment of the second invention of this application, FIG. 4 is a block diagram for explaining the prior art of this application, and FIG. 5 is a diagram explaining the operation of the prior art. 6 to 9 are front views for explaining types of color unevenness. 10: color image data source, 20: conversion means, 30A: hue image memory, 30B: lightness image memory, 30C: saturation image memory, 31:
Hue value reading means, 32: color correction value calculation means, 33: correction means,
34: average saturation calculation means, 35: maximum value storage means, 50: adder, 51: determination means.

Claims (2)

(57) [Claims] 1. A process for converting color image data into saturation image data and hue image data; and B. a process for subdividing the saturation image data into a plurality of local blocks consisting of m × n pixels. The saturation of each pixel constituting a plurality of local blocks is corrected by a visual correction coefficient determined by a hue value in the hue image data at the same pixel position, and the average value of the corrected saturation is calculated for each local block. The process of obtaining, C. the process of selecting and storing the maximum value from the average saturation of each local block, and D. changing the number of pixels constituting the local block, and calculating the average saturation of each local block. The process of selecting and storing the maximum value from the average saturation of each local block in the same manner as described above, and E. Changing the number of pixels constituting the local block a plurality of times, and storing for each pixel number Of the maximum values of the average saturation And F. a process of determining that there is color unevenness when the value of the sum is equal to or greater than a predetermined value. 2. A conversion means for generating hue image data and saturation image data from color image data; and B. a hue image memory and a saturation image memory for storing these hue image data and saturation image data. And C. subdividing the saturation image data stored in the saturation image memory into a plurality of local blocks consisting of m × n pixels,
A hue value reading means for reading from the hue image memory a hue value at the same pixel position as each pixel in each subdivided local block; and D. calculating a color correction coefficient based on the hue value read by the hue value reading means. A color correction coefficient calculating means; E. correcting means for correcting the saturation of pixels constituting the subdivided local block by the color correction coefficient calculated by the color correction coefficient calculating means; F. correction by the correcting means Means for calculating the average value of the saturation of the obtained values for each local block; and G. selecting and storing the maximum value of the average saturation of each local block calculated by the average saturation calculating means. Maximum value storage means, H. Change the number of pixels constituting the local block, and let the average saturation calculation means calculate the average saturation in the local block in which the number of pixels has been changed, and change the number of pixels a plurality of times. Control hand that performs repetitive control If, I. a color image signal evaluating apparatus constructed and adding means for obtaining the sum of the maximum value stored in the maximum value storing means for each change of the number of pixels constituting the local block, by.