JP2948291B2 - Evaluation device for color image signal - Google Patents

Abstract

PURPOSE:To evaluate every kind of color shading of a color image signal from a color image pickup element quantitatively by detecting every kind of color shading by making into numeric values by plural color shading evaluation devices, finding the sum of the values after standardizing by multipying the numeric value of detected color shading by a coefficient, and deciding the normal/defective condition of the color shading based on whether the value of the sum exceeds or less than a prescribed value. CONSTITUTION:Color image data R, G, and B read out from a color image data source 10 are supplied to an HLS conversion part 20, and are converted to a hue image data H, lightness image data L, and saturation image data S, and are stored in hue data memory 30H, lightness data memory 30L, and saturation data memory 30S, respectively. An arithmetic part 31 fractionizes the saturation image data S in the saturation data memory 30S into blocks KB consisting of picture elements, and reads out the saturation image data S at every block, and calculates mean saturation in each block KB, and extracts and stores the maximum value of the mean saturation of the block KB. Also, the part calculates the sum of all obtained maximum mean saturation, and outputs the sum as fundamental coloring degree Ts. Thereby, it is possible to evaluate every kind of color shading of the color image from the color image pickup element quantitatively.

Description

The present invention relates to a color image signal evaluation apparatus using a method for quantitatively evaluating a color image signal obtained from a color image sensor such as a CCD made of a semiconductor IC. About.

"Prior art" In recent years, a color imaging device in which a mosaic color filter or a stripe color filter is mounted on the front surface of a CCD imaging device made of a semiconductor to obtain a color image signal has been put to practical use. . CCDs are made by semiconductor IC technology, but defects occur due to various causes in the manufacturing process. Furthermore, when a color filter is printed on a CCD to produce a color image sensor, an alignment error occurs. When a color image signal obtained from such an image sensor is reproduced, various color unevenness occurs due to a defect or misalignment of the image sensor. For example, when light of any single color without a pattern, for example, white light, is uniformly applied to the entire image receiving surface of the image pickup device, (1) colored stripes are formed diagonally on the playback screen, or (2) vertical stripes appear on the playback screen. Alternatively, colored stripes occur in the horizontal direction, or (3) a large area in the reproduction screen is colored.

[Problems to be Solved by the Invention] Conventionally, a color image signal from a color image sensor is projected on a color cathode-ray tube, and the screen is visually monitored to determine whether the above-mentioned color unevenness is less than a predetermined allowable level. , The efficiency is low. In particular, a large number of inspectors are required for quality inspection of a manufactured image sensor in a mass production factory of the image sensor. In addition, there is an inevitable difference between individuals in the evaluation criteria for various color unevenness among the inspectors, and as a result, there is a problem that the quality of the color unevenness of a product that has passed the inspection has a large variation.

[Means for Solving the Problems] An object of the present invention is to provide a color image signal evaluation apparatus using a method for quantitatively evaluating various color unevenness of a color image signal from a color image sensor.

According to the present invention, a total sum is obtained by multiplying output evaluation values of a plurality of color unevennesses based on different viewpoints by respective determined coefficients. The quality of the color image signal is determined by comparing the sum with a reference value.

According to the first aspect of the present invention, a color image signal is converted into chroma image data, and one screen is stored in a chroma data memory. The screen is divided into blocks of m × n pixels, the saturation data in the saturation data memory is read, the saturation is averaged for each block, and the maximum value of the average saturation is obtained. A similar process of increasing (or decreasing) the block size to obtain the maximum average saturation is repeated a predetermined number of times, and a value related to the sum of the maximum average saturations is detected as the basic coloring degree.

According to the second aspect of the present invention, the color image signal is converted not only into saturation data but also into hue data, and the hue data is stored in a hue data memory. In the process of detecting the basic coloring degree, the hue data is read from the pixel position in the hue data memory corresponding to the read saturation data, and the saturation data is corrected based on the hue data. The basic coloring degree is obtained in the same manner using the data.

According to the third aspect of the present invention, the color image signal is converted into hue data and saturation data, and stored in the hue data memory and the saturation data memory, respectively. The saturation data at each pixel position is read from the saturation image memory, the hue data at the pixel position having saturation greater than the reference saturation is read from the hue image memory, and the color represented by the hue data is identified to identify the corresponding color. "1" is written to the corresponding pixel position of the matrix memory. A correlation coefficient is calculated for the X and Y coordinates of the pixel position where "1" is written in each color identification matrix memory, and a value corresponding to the correlation coefficient is obtained as a diagonal stripe pattern degree.

According to the fourth aspect of the present invention, the color image signal is converted into hue data and saturation data, and stored in the hue data memory and the saturation data memory, respectively. The saturation data at each pixel position is read from the saturation data memory, the hue data at the pixel position having saturation greater than the reference saturation is read from the hue data memory, and the color represented by the hue data is identified to identify the corresponding color. At the same time, "1" is written to the corresponding pixel position in the matrix memory, and the saturation data is stored in the corresponding pixel position in the saturation image memory of the corresponding color. The stored contents of each color identification matrix memory are integrated in the Y direction and the X direction, respectively, to obtain an X integration count and a Y integration count, and these are weighted to be large at the peripheral portion of the screen. On the other hand, the saturation data stored in the saturation image memory of each color is integrated in the Y direction and the X direction to obtain X and Y integrated saturation, and each X and Y in which the weighted X, Y integration count becomes maximum is obtained. The position where the integral saturation exceeds the reference integral saturation on both sides of the Y coordinate is defined as a stripe boundary. Calculate the sum of the saturation within the width of the stripe and the standard deviation of the saturation at the periphery and add the integral count within the width of the stripe. A value corresponding to the relationship between the sum of the chroma, the standard deviation, and the sum of the integral count is obtained as the peripheral stripe pattern degree.

According to the fifth aspect of the present invention, the color image signal is converted into hue data and saturation data, and stored in the hue data memory and the saturation data memory, respectively. The saturation data at each pixel position is read from the saturation data memory, the hue data at the pixel position having saturation greater than the reference saturation is read from the hue data memory, and the color represented by the hue data is identified to identify the corresponding color. At the same time, "1" is written into the corresponding pixel position in the matrix memory, and at the same time, the value corresponding to the saturation data is stored in the corresponding pixel position in the saturation image memory of the corresponding color. The contents stored in each color identification matrix memory are integrated in the Y direction and the X direction, respectively, to obtain an X integration count and a Y integration count, and these are weighted to be large at the center of the screen.
On the other hand, the values stored in the saturation image memory of each color are integrated in the Y direction and the X direction to obtain X and Y integral saturations, and each X, Y coordinate at which the weighted X, Y integral count becomes maximum The positions where the integrated saturation exceeds the reference integrated saturation on both sides of are defined as stripe boundaries. The sum of the saturation within the stripe width and the standard deviation of the saturation at the center are calculated, and the standard deviation of the position of the pixel having "1" at the center is calculated. A value corresponding to the result obtained by dividing the product of the sum of the saturation and the standard deviation of the saturation by the standard deviation of the position is obtained as the center stripe pattern degree.

According to the invention of claim 6, the basic coloring degree, the oblique stripe pattern degree, the peripheral stripe pattern degree and the central stripe pattern degree obtained based on the inventions of claim 1 to claim 5 have predetermined coefficients respectively. The multiplication is performed, and the sum of the multiplication results is compared with a reference value to obtain an evaluation result of the color image signal.

[Embodiment] [Invention of Claim 1] FIG. 1 shows a color image signal evaluation apparatus for evaluating a basic color degree, which is a color image signal evaluation method according to the invention of claim 1. The basic coloration degree is such that when standard white light with no pattern is incident on the entire surface of the image received by the image sensor, the color formed on the display surface of the color CRT by the output R, G, and B signals from the image sensor as a whole It is a scale for evaluating color unevenness that is colored to a very thin color partially or partially.

In FIG. 1, reference numeral 10 denotes a color image signal source obtained from the output of a color image sensor such as a color CCD. The color image signal source 10 may be output terminals 10R, 10G, and 10B in which R, G, and B signals output from the color image sensor are converted into digital signals by an AD converter, or stores one frame of the digital signals. Memory, 10R, 10G, 10B. Here, the latter case will be described. The three memories 10R, 10G, and 10B store, for example, a color image signal obtained by uniformly entering a non-patterned white light on the image receiving surface of the color image sensor, a single color signal of red, green, and blue (hereinafter simply referred to as R, G , B signals), and stores the R, G, B image signal data obtained by AD-converting the separated single-color signals.

The color image data R, G, and B read from the color image data source 10 are given to an HLS conversion unit 20. The HLS conversion unit 20 converts the hue image data H, lightness image data L, and saturation image data S into data. Convert.

The hue image data H is represented by a numerical angle from 0 to 360. For example, red is 0 ~ 30 ゜ and 330 ~ 360 ゜, yellow is 30 ~
90 ゜, green is 90 ~ 150 ゜, cyan is 150 ~ 210 ゜, blue is 210 ~
270 ゜, magenta is 270-330 ゜. In addition, each color may be classified in the range of a 12-bit number.

The lightness represents the brightness of the display image, and the saturation represents the vividness of the color, each represented by a value in the range of 0 to 1.0.
Algorithms for calculating the hue H, lightness L, and saturation S of a color synthesized by these signals from given R, G, B signal levels, that is, an HLS conversion algorithm are well known.
The description is omitted. Approximately equivalent conversion such as HSV conversion may be used instead of HLS conversion.

The hue image data H for one screen, the brightness image data L, and the saturation image data S obtained by conversion in the HLS conversion unit 20 are stored in a hue data memory 30H, a brightness data memory 30L, and a saturation data memory 30S, respectively. It is memorized.

In the evaluation of the basic coloring degree according to the invention of claim 1,
Since only the saturation image data S among the image data H, L, S stored in the data memories 30H, 30L, 30S is used, the hue data memory 30H and the brightness data memory 30L are unnecessary.

Calculating unit 31 is subdivided into blocks KB comprising the saturation image data S in the chroma data memory 30S from the pixels of the m × n points in First Step S ₁ as shown in the process flow of FIG. 2, step
Reads the saturation image data S for respective blocks S _2, it calculates the average saturation in each block KB. Then it is determined whether reading of all the blocks has been completed for entire field in step S _3,
If you have not finished the process returns to step S _2. In this way, the block KB performs this processing over the entire screen without overlapping each other. The number of pixels m × n of the block KB is, for example, about 10 × 10 as an initial value.

When this process is performed over the entire screen, step S ₄
Extracts and stores the maximum value of the average saturation of the block KB.
The size of the block KB in step S _5, it is determined whether or exceeds a predetermined magnitude, beyond otherwise be blocked KB in step S ₆
Is changed to m = m + k ₁ , n = n + k ₂ and the process returns to step S _{1. In} steps S ₂ and S ₃ , the average value of each block KB is obtained over the entire screen again. _{In step 4} , the maximum value of the average saturation of each block KB is extracted.

Calculating an average color saturation of each block KB over the entire screen each time it changes the number of pixels of the block KB in step S ₁ to S ₅ as described above, extracts the maximum value of their average saturation,
Repeat to remember.

When block KB is sized to be able to take four for example in the screen in step S _5, the sum of all maximum average saturation obtained so far calculated in step S _7, the basic coloring the sum Output as degree T _S.

Since the size of the block KB gradually changes in this way,
For example, one local color unevenness gives a certain large maximum average saturation value until the block size becomes about the size of the color unevenness, and includes a color unevenness when the block size exceeds the size of the color unevenness. The average saturation of the blocks becomes progressively smaller. On the other hand, the saturation over the entire screen gives a substantially constant maximum average saturation according to the intensity of the saturation regardless of the change in the block size. Therefore, it can be determined that the smaller the value of the basic coloring degree T _S (sum of the maximum average saturation) is, the smaller the color unevenness and the intensity of the color image signal are.

FIG. 3 shows a graph in which the maximum average chroma value of the block KB in each block size is plotted based on the actually measured data. Step S obtained in ₆ basic color degree T _S (the sum of the maximum mean saturation) represents an area surrounding of these curves A~F and coordinate the horizontal axis. As shown by curves A, B, and C shown in this graph, when the characteristics exhibit a characteristic close to a monotonically decreasing function in which the maximum average saturation sharply decreases as the block size m or n increases, the products are non-defective, and the curves D and E are obtained. If the function does not become a monotonically decreasing function as in F and F, it tends to be defective.

As described above, according to the first aspect of the present invention, the chroma image data is subdivided into blocks KB, the average chroma of each block is calculated, and the maximum value of the average chroma is calculated while changing the block size. It is extracted for each block size, the maximum value of the extracted average saturation is cumulatively added, and this cumulative added value is output as the basic coloring degree. The basic coloring degree is compared with the quality judgment criterion. If the standard coloring degree is smaller than the quality judgment criterion, it is judged as good, and if it is larger than the quality judgment criterion, it is judged as defective. Therefore, a more reliable inspection can be performed as compared with the case where the inspection is performed visually. Further, the inspection can be performed automatically. As a result, a large number of test objects can be inspected in a short time, and labor saving can be achieved.

In the above description, the processing is terminated when the size of the block becomes 1/4 in the screen. However, the number is not necessarily limited, and the number can be arbitrarily searched. Although the blocks have been described as not overlapping each other, the same result can be obtained by partially overlapping the blocks.

Further, it can be easily understood that the present invention can be used as a method and an apparatus for evaluating a color image signal without being limited to a test of a color image pickup device made of a semiconductor.

[Invention of Claim 2] In the invention of Claim 1 described above, the basic coloring degree is defined by focusing only on the saturation value so that the presence or absence of color unevenness can be determined. The color reproducibility varies due to the difference in sensitivity and the difference in the characteristics of color filters provided on the front of the image sensor (differences in filter types), resulting in differences in color image sensor models. For example, there is a phenomenon that a certain color is emphasized or lower than other colors.

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

In the second aspect of the present invention, the basic coloring degree which has solved the disadvantage is detected as a color unevenness evaluation scale. For this reason, the hue data read out from the hue data memory 30H is taken into the arithmetic unit as shown by a broken line in FIG. 1 by using the hue data memory 30H of FIG. 1 which is not used in the invention of claim 1.

Chroma data memory 30 in step S ₂ of the processing flow shown in FIG. 2 by the calculation unit 31 performs the invention of claim 2
When obtaining the average saturation in each block KB in S, the hue value at the pixel position corresponding to each pixel constituting this block KB is read from the hue data memory 30H, and the color correction value is calculated from these hue values. The color saturation values of the pixels constituting each block KB are calculated and corrected by the color correction values, and the corrected saturation values are averaged to calculate the average saturation of each block KB. FIG. 4 shows the processing flow. Steps
M for each block KB consists chroma data memory 30S at S ₂₁
At the same time as reading the saturation values of × n pixels, the hue value H is read from the pixel position in the hue data memory 30H corresponding to each pixel. Obtaining the color correction value F in this hue values H step S _22.

Calculating color correction value F in step S ₂₂ is determined, for example, from the following equation.

Ao, Ai, and Bi are coefficients.

As an example, h = 2, Ao = 2.0, A ₁ = 1.0, A ₂ = 0.5, B ₁
FIG. 5 shows the characteristics of the color correction value F when B = 300 and B ₂ = 300. By appropriately setting the values of h, Ao, Ai, and Bi, the correction characteristics shown in FIG. Can be set.

The color correction value F is multiplied by the saturation image data S in step S ₂₃ to obtain the corrected chroma. Using the corrected chroma value in step S ₂₄ calculates the average saturation of each block KB.

The number of pixels m × n is the same as in the processing of steps S ₂ and S _{3 in} FIG.
When the calculation of the average saturation of the block KB in size over the whole screen, selected maximum value among them at the step S _4,
The maximum value of these average saturation obtained by through step S ₁ to S ₅ is added in step S _7. Note that it may be cumulatively added for each obtain maximum average saturation at Step S ₄ instead of adding all of the maximum average saturation at a time, in step S ₇ is obvious.

The basic coloring degree Ts, which is the addition result obtained in this way, is compared with a reference value, and if it exceeds the reference value, it can be determined to be defective, and if it is less than the reference value, it can be determined to be good.

As described above, according to the second aspect of the present invention, when calculating the average saturation, the saturation characteristic of each pixel is corrected by the color correction value obtained from the hue value at the pixel position, thereby obtaining the sensitivity characteristic to the color of the eyes. Alternatively, it is possible to determine whether the color filter is suitable for the characteristics of the color filter provided on the front surface of the image sensor.

In particular, since the color correction value F can be set to an arbitrary value for each color as shown by the curve in FIG. 5, for example, by appropriately setting the value of h shown in the equation (1), the color correction value F is output from the test object. 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 determined to be good, and the color unevenness appears only faintly. If it appears to be emphasized to the eyes, it can be determined to be defective,
For example, each CCD manufacturer can classify according to its own criteria.

Although the example in which the color correction value is calculated for each pixel has been described above, the same block KB is also taken in the hue data memory 30H corresponding to the block KB for which the average saturation is calculated, and the block KB is calculated. The color correction value F may be obtained from the average saturation of Equation (1) using Equation (1).

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.

[Invention of Claim 3] FIG. 6 shows a color image signal evaluation apparatus for evaluating oblique stripes, which is one of the color image signal evaluation methods according to the invention of claim 3.

As in the case of FIG. 1, 10 shown in FIG.
2 shows a color image signal source which is an output of a color image sensor such as a CCD. A non-pattern, for example, a white screen image is formed on the entire imaging surface of the color imaging device. As a result, the red signal R, the green signal G, and the blue signal B are separately output from the color image signal source 10.

Each monochrome signal R, G, B output from the color image signal source 10
Are converted into hue image data H and saturation image data S by the HLS converter 20.

The hue image data H and the saturation image data S for one screen obtained by the HLS conversion unit 20 are stored in a hue data memory, respectively.
30H and are taken into the saturation data memory 30S. That is, assuming that the total number of pixels in one screen is N, N pieces of hue image data H and saturation image data S are obtained. The figure shows a case in which the brightness image data L is also obtained. The brightness image data L is stored in the brightness data memory 30L.

Data to be evaluated with respect to the saturation image data S in the saturation data memory 30S is limited to the saturation image data S having a reference value or more. The saturation reference value may be an average saturation value per pixel of the saturation image data S stored in the saturation data memory 30S, or a saturation value obtained by multiplying the average saturation by a constant coefficient. 0 regardless of the saturation image data S
The above-mentioned predetermined constant value may be used. In the case of 0, all the saturation image data S are to be evaluated. In this second viewpoint, the calculation unit 31 determines the reference value of the saturation with the highest frequency according to the processing flow of FIG. 7, and calculates the saturation coefficient based on the correlation coefficient of the pixel position of the saturation equal to or higher than the reference value. To determine the degree of oblique stripe pattern.

First find reading histogram saturation image data S from the chroma data memory 30S in a step S _1, to obtain the value of the most frequent color saturation. FIG. 8 shows an example of a histogram of the saturation image data S. This histogram is obtained by equally dividing the width of saturation 0 to 1.0 into a fixed number, for example, 1,000 sections, and calculating the number of pixels belonging to each saturation section. The center saturation of each saturation section is set as the saturation representing the section. In this example illustrates the case shown the most frequent value (referred to as mode value) in the saturation value S _m.

Then a reference saturation value p · S _m multiplied by the predetermined ratio p in this mode saturation value S _m in step S _2. In this example, a description will be given assuming that the predetermined ratio p = 0.5.

Then extracts pixels having a saturation higher than the reference saturation value 0.5S _m from chroma data memory 30S at Step S _3. Further, the hue image data H is read from the hue data memory 30H by using an address having the reference saturation or higher, the color of the hue image data H is identified by the color identification circuit 32, and the color identification matrix memory 40R to 40MK corresponds to the corresponding memory. Write "1" to the address (pixel position). “0” is held in addresses corresponding to other pixel positions. Therefore, the color identification matrix 40R~40M "1" to the pixel position with the saturation of more than 0.5S _m for each color is stored. Color identification matrix 40R,
The colors corresponding to 40G, 40B, 40Y, 40C, and 40M are red for 40R, green for 40G, blue for 40B, yellow for 40Y, cyan for 40C, and magenta for 40M.

Then the X-coordinate of the pixel position is "1" in the color identification matrix in step S _4, the correlation coefficient r for each color with respect to the Y-coordinate, i.e. in this example respectively obtained for the six colors. The correlation coefficient r is given by the following equation.

Here, (x _i , y _i ) is a pixel position where “1” is stored, n is the number of pixels, and
x _i , y _i (i = 1,..., n) are average values.

The absolute values of the correlation coefficients r for these six colors are taken, and those values are set to R. Each color identification matrix 40
R takes a large value when pixel positions where "1" is stored in R to 40M are distributed so as to form an oblique pattern as shown in FIG. On the other hand, when the pixel positions where “1” is stored are dispersed as shown in FIG. 10, R has a small value.

Step S at ₅ in the color identification matrix 40R～40M, the total number of pixels of "1" is written image region pixel number n was N
To obtain a value K = n / N in which n is normalized. This normalization is performed so that color evaluations can be compared even for imaging elements having different numbers of pixels.

In Step S ₆ for each color calculated R and K for each color determining the I = R × K. This I is defined as the diagonal pattern degree of each color. Calculation unit 31 further includes at step S ₇ performs the above processing obtains the largest of the diagonal pattern of I of each of these colors, and representative oblique pattern of I _m of the color image signal the maximum oblique pattern of I do. Or all of the average values of the oblique pattern of the the color may be selected as the representative oblique pattern of I _m. Calculating section 31 outputs a representative oblique pattern of I _m determined in this way. For example, when the representative diagonal pattern degree _Im is equal to or more than a predetermined value, it can be determined that the image sensor is defective.

May be performed with weight on color-coded to I = R × K hitting the calculation of each color in the diagonal pattern of I in step S _6. That is, there are conspicuous colors and inconspicuous colors. For this reason, it is conceivable to multiply the oblique pattern degree I of a conspicuous color by a large weighting coefficient W to make the judgment criteria different for each color. As the weighting coefficient W, for example, W = 1.0 for red R and magenta M, W = 0.75 for green G and yellow Y, and W = 0.5 for cyan C and blue B.

It is possible to increase the detection sensitivity with respect to the oblique pattern of color that stands out by defining the swash pattern of I to the result of multiplying the weighting coefficient W maximum or an average value representative oblique pattern of I _m for each color.

As described above, according to the third aspect of the invention, it is possible to mechanically detect color unevenness occurring as an oblique pattern included in a color image signal, not artificially. Therefore, the color image signal from the color image sensor can be evaluated in a short time. Therefore, a large number of color image elements can be tested in a short time, and labor saving can be achieved.

In addition, since uniform evaluation can be performed, there is obtained an advantage that highly reliable inspection can be performed.

In the above description, six colors of red, green, blue, yellow, cyan, and magenta are used. However, three to twelve colors of red, green, and blue can be processed separately and are limited to six colors. Not easily understandable.

[Invention of Claim 4] Next, a color image signal evaluation apparatus for evaluating vertical and horizontal stripes around a screen by the color image signal evaluation method according to the invention of claim 4 will be described with reference to FIGS. No.
As shown in FIG. 11, the configuration is such that the arithmetic unit 31 in the apparatus of FIG. 6 is further provided with a saturation image memory 50R corresponding to each color.
M50M are connected. Also in this evaluation method, non-patterned white light, for example, is incident on the entire image receiving surface of the image sensor (not shown). The red, green, and blue monochromatic signals R, G, and B from the color image signal source 10, which are the outputs of the image sensor, are converted into the hue image data H and the brightness image data in the HLS conversion unit 20 as in the case of FIG. L and saturation image data S are stored in the hue data memory 30H, the brightness data memory 30L, and the saturation data memory 30S, respectively.

Figure 6 when the saturation image data S obtained in the saturation data memory 30S similarly obtain the value S _m of saturation of the most frequent portions a histogram of saturation is read out to the arithmetic unit 31 . The reference saturation is determined from the mode saturation value Sm in the same manner as in the case of FIG. 6, and the value of the saturation image data S in the saturation data memory 30S is the position of a pixel having a value equal to or greater than the reference saturation. That is, the address is given to the color identification unit 32. Color identification section
32 identifies each color of the hue image data H in the memory 30H corresponding to those addresses, selects the color identification matrix memories 40R to 40M of the corresponding colors, and sets "1" to the corresponding address. Writing is the same as in FIG. As a result, red color image data in which the color saturation image data S has a value equal to or higher than the reference color saturation is stored in the color identification matrix memory 40R.

Hereinafter, the same applies to the color identification matrix memories 40G to 40M.

The arithmetic unit 31 further executes the following processing according to the flow shown in FIG. 12 using the data stored in the saturation data memory 30S and the color identification matrix memories 40R to 40M.

Step S ₁ : The size of the image is represented by the number of pixels N _X and N _y in the X-axis direction and the Y-axis direction. Therefore N = N _x × N _y. In the invention of claim 4, an X register 40x and a Y register 40y are provided in association with each of the color identification matrix memories 40R to 40M as shown in FIG. The operation unit 31 integrates the number of pixels of 1 stored in each column and each row in each color identification matrix in the Y and X directions, respectively, and obtains N _X integration results C _X
Obtain N _y integration results _Cy .

These integration results are stored as C _X and C _y in the X register 40x and the Y register 40y. These integrated values C _X , _Cy are hereinafter referred to as integral counts. Each N _X number of integral count C _X As evaluation of stripes is equal to the X and Y directions multiplying N _X / N _y. Or it may be multiplied by N _X / N _y respectively integral count C _y. In the following, for the X axis, the integration count thus standardized is used as _CX . The above processing is performed for the color identification matrix memories 40R to 40R for each color.
Performed for 0M.

Step S ₂ : In the invention of claim 4, in order to detect and evaluate fringes at the peripheral portion of the screen, the integral count C _x (normalized) taken into the X register 40x and the Y register 40y of each color. and with respect to C _y, performs weighted shown in Figure 14, respectively. For this weighting, for example, a 1/10 portion of the length from both ends of each of the X register 40x and the Y register 40y is weighted by a part of a cos function called an inverted Tukey 80% window, for example. This weighting gives detection sensitivity to the peripheral portion of the screen.

Step S _3: arithmetic unit 31 for each color integration count C _X, C _y
Location X _m but with the maximum value, respectively, obtains the Y _m.

Step S _4: saturation data memory 30S saturation image data S of the pixel position value of the further arithmetic unit 31 in each color identification matrix memory 40R~40M is 1 in the invention as claimed in claim 4
From the corresponding color image memory 50R to 50M. An X register 50x and a Y register 50y are provided in association with each of the saturation image memories 50R to 50M, and the saturation values in the memories are respectively integrated in the Y direction and the X direction to obtain Nx integration results S _X And N _y integration results S _y are obtained. These integration results will be referred to as integral saturation.

The integrated saturations S _X and S _y are stored in an X register 50x and a Y register 50y. The integral saturation S _X is normalized by multiplying N _x / N _y in the same manner as the integral count C _X. Hereinafter used on a normalized for S _X. The above processing is executed for the color saturation image memories 50R to 50M.

Step S _5: the position X _m of integrating count C _X, C _y is the maximum, respectively for each color of the color identification matrix memory 40R～40M,
Obtain the corresponding integral saturation S _X , S _y values at Y _m as S _Xm , S _ym .

Step S ₆ : Position where the value of integral saturation of each color is S _Xm , S _ym
Integral saturation positions A and B, which are, for example, 0.2 times the values of S _Xm and S _ym , respectively, are detected toward both ends in the X and Y directions with X _m and Y _m as centers. FIG. 15 shows an example of the integral saturation S _X in the X register 50x relating to one saturation image memory.
It shows about.

When a position at which the magnification becomes 0.2 times cannot be detected on either side of each of X _m and Y _m , that is, when the integrated saturation S _X and S _y are all 0.2 on that side.
S _Xm, in the case of greater than 0.2 S _ym to the end of the X register 50x or the Y register 50y and its location. Processing in step S ₆ defines the end positions in the width direction of the vertical or horizontal stripes are trying to detect.

Note this position A, step S each X register 50x before performing the process of _6, Y register integral saturation S _X held in 5Oy, length to S _y 3 determine a B (i.e. 3 pixels length) Is preferably applied to reduce the noise. The low-pass filter having a length of 3, for example, in the case of the X register 50x, indicates a low-pass characteristic obtained by the following processing.

That is, the data fetched into the X register 50x for one color is S _X1 , S _X2 , S _X3 ,..., S _Xi,.
(I = 1,..., N _x ), the successive data, for example, three data S _{X (i−1)} , S _Xi , S _{x (i + 1)} are respectively multiplied by 1
The weighting coefficients W ₁ , W ₂ , and W ₃ are multiplied by ₂ , 13, and 14, and are then given to the averaging means 11 to add up the multiplication results. For example weighting coefficients _{W 1 = 0.25, W 2 =} 0.5, the output _Xi by the following equation _Xi = 0.25 S _X with averaging means 11 is _{W 3 = 0.25 (i-1} ) + 0.5S Xi + 0.25S X ( _{i + 1)} ... (6) This value _Xi i = 1, ..., it is calculated for N _X. However, S _X0 = 0 and S _{X (Nx + 1)} = 0. These results are sequentially stored in the auxiliary register 50a. If _Xi is obtained in the register 50a for all i, rewritten register 50 _X by its contents. By this filtering, a sudden change in the value of the adjacent data S _Xi , S _{X (i + 1)} is made gentle. The same processing for the data S _y in the Y register 5Oy.

Step S _7: Step S colors of the integration saturation S by treatment with ₆
The distance between two positions A and B detected as shown in FIG. 15 (only shown for S _X ) for each of _X and S _y ,
That is, the widths of the vertical stripes and the horizontal stripes are D _X and D _y , respectively.

Step S ₈ : 2 for each of the integral saturations S _X and S _y of each color
The integral saturation S _X , S _y between the two positions A, B is added, and the addition result is V _X , V _y . However, when D _X = 0, V _X = 0, and when D _y = 0, V _y = 0. These values are stored in each saturation image memory 50R.
It means the integrated value of the saturation S of the sample pixel on the stripe having the width defined by A and B at ~ 50M.

Step S ₉ = integral count C _X between two corresponding positions A, B for each of the integration count C _X, C _y of each color, by adding the C _y, is the addition result K _X, and K _Y. However, when D _X = 0, K _X = 0, and when D _y = 0, K _y = 0. These values are A, B
Means the area of the stripe having the width defined by

Step S _10: defining next given by Equation alpha _v, vertical stripes parameter alpha _h respectively, and horizontal stripes parameters In equation (7), (V _X ÷ D _X ) represents the average integrated saturation per pixel in one column on the vertical stripe of width D _X , and is further divided by the size N _X of the screen in the X direction. Is becoming Normalized value is weighted by further weighting function W _X that is determined by D _X / N _X. The same applies to equation (8).
Note that J is selected to be, for example, about J = 5.

Step S _11: the weight Since the human eye is not a bright thinner than a certain degree in stripes negligible stripe width D _X, D _y Whereas parameters K (e.g. K = 10) or less thin stripes, the further the following formula Multiply.

α _v ′ = α _v × {0.5 [1-cos (2π · Dx / 2K)]} ^L (9) α _h ′ = α _h × {0.5 [1-cos (2π · Dy / 2K)]} ^L (10) L is a value of, for example, about 0.67. By appropriately selecting the weight L, α _v ′ and α _h ′ can approach zero at a desired speed as the width of the stripe becomes narrower than K.

Step S _12: the aforementioned respective color identification matrix memory 40R~40M
, The pixel positions X _j , Y _{k of} all the pixels in which “1” within the previously defined stripe width is stored in the color identification matrix memory 40 for the maximum integration count positions X _m , Y _m. for the standard deviation d _X, obtaining the d _y.

 However, this is obtained by the following formula using the modified standard deviation.

S _j, S _k is the saturation of each sample pixel corresponding to _{X j, X k, n x} , is n _y is the number of samples pixels having "1" in the prescribed in stripe width D _x D _y .

Step ₁₃ : Calculate the following equation for each color.

_{Q X = α X × d X} × K X ... (13) Q y = α y × d y × K y ... (14) Step S _14: the calculation result in step S ₁₃ is 6 pairs obtained a total of 12. Let the maximum value among them be Qmax.

Step S _15: Step S ₉ obtained in six sets of K _X, adding all the K _y is a Z.

Step S ₁₆ : P _m = Qmax × Z The calculation of (15) is performed, and this P _m is set as the peripheral stripe pattern degree.
Calculating section 31 outputs the periphery fringe pattern of P _m thus calculated.

Peripheral fringe pattern degree obtained as a result of the processing described above
The magnitude of the P _m can be performed to evaluate the color image signal. That can be vertical or horizontal stripes by more irregular color large value near portion fringe pattern of P _m is determined to be occurring on the periphery of the screen. Therefore, the quality of the image sensor can be determined based on whether the value exceeds a predetermined value.

As described above, according to the fourth aspect of the invention, it is possible to quantitatively detect a striped pattern due to color unevenness occurring in the peripheral portion of the screen. As a result, a more uniform inspection can be performed as compared with the case of visual inspection. Of course, since the processing for calculating the degree of pattern is performed using a computer, the inspection can be performed automatically, and the labor can be saved.

In the embodiment described above, the curve giving the weighting function is represented by cos
Although a function is used, a polygonal line function can also be used. It can be easily understood that the length of the low-pass filter is not limited to three.

[Embodiment 5] Next, referring to FIGS. 17 to 20, a color image signal evaluation apparatus for evaluating vertical and horizontal stripes at the center of an image by a color image signal evaluation method according to a fourth aspect of the present invention will be described. Show. The configuration of this device is basically the same as the device of FIG. 11 except that a histogram flattening unit 33 is added. Also in this evaluation method, for example, white light is made incident on the entire image receiving surface of an image sensor (not shown),
Let the R, G, B signals be the R, G, B signals from the signal source 10 in FIG.

The processing of processing the obtained R, G, B signals and writing "1" to the corresponding pixel position of the color identification matrix memory 40R to 40M of the corresponding color for the pixel having a color equal to or greater than the reference saturation value is also described. The description is omitted because it is the same as in FIGS. 6 and 11.

According to the fifth aspect of the present invention, the chroma histogram obtained by the arithmetic unit 31 is supplied to the histogram flattening unit 33, and the histogram is flattened. The flattening process indicates that the process is performed such that the average number of pixels per unit saturation is constant over the entire screen. This histogram flattening process enhances the contrast of the color vividness.

The saturation image data emphasized by the histogram flattening unit 33 is stored in the saturation image memories 50R to 50R as in the case of FIG.
Written to 0M. Hereinafter, the processing performed by the arithmetic unit 31 will be described with reference to the processing flow of FIG.

Step S _1: Figure 11 the same manner as in the case of each color identification matrix memory 40R~40M the X and Y directions on the integral-obtained integrated count value C _x, the same register as showed C _y in Figure 6 40x
And take in to 40y. The same applies to multiplying the integration count C _X by the image size ratio N _x / N _y , thereby normalizing and removing the effect of the difference in image size.

Step S _2: To evaluate and detect fringes in the central portion of the screen in the invention as claimed in claim 5, in the X register 40x and the Y register 40y on each color of the identification matrix memory 40R～40M (see FIG. 13) Integral counts C _X , _{Cy taken}
Are weighted as shown in FIG. This weighting is, for example, 1 in length from both ends of each register 40x, 40y.
The / 10 part is called, for example, Tukey's 80% window
Weighting is performed by a part of the function of s, and detection sensitivity is given to the center of the screen.

Step S _3: location X _m the integration count C _X for each color identification matrix memory 40R~40M, C _y is the maximum value, respectively,
Find Y _m .

Step S _4: integrating the saturation value after the histogram flattening stored in the saturation image memory into registers 50x and 50y provided in association with each saturation image memory 50R~50M in Y and X direction As a result, S _X and S _y are stored. The integral saturation S _X stored in the register 50x is multiplied by N _X / N _y and normalized to eliminate the influence of the screen size, as in the case of FIG.

Step S _5: the corresponding at position X _m, Y _m where Figure 12 step S ₅ similarly to the color identification matrix according to the memory 40R~40M registers 40x, integrated count C _x in 40y, C _y is the maximum, respectively The values of the integral saturations S _X and S _y are obtained as S _Xm and S _ym . However, the integral saturation of each color and each direction is weighted as shown in FIG. The curve for performing the weighting is the length of each of the chroma image memories 50R to 50M in the X and Y directions, that is, the X and Y of the screen.
This is a modification of the cos function in which the directional sizes N _x and N _y are each one cycle.

Integral saturation stored in registers 50x and 50y for each color
S _X and S _y are subjected to a low-pass filter having a length of, for example, 3 as described with reference to FIG. 16 to reduce noise.

Step S _6: Figure 12 step S ₆ similarly to the color of the integral saturation S _X, the value S _Xm of S _y, S _ym a is position X _m, toward the screen across the around the Y _m S _Xm and _Sym values, for example, 0.5
The positions A and B of the doubled integral saturation are detected (see FIG. 15). When it is not possible to detect a position that becomes 0.5 times on either side of each of X _m and Y _m , that is, when the integral saturation S _X , S _y is larger than 0.5S _Xm , 0.5S _ym on that side, register 50x and 50
Let the end of y be that position.

Step S _7: two detected positions A for each color, the distance or vertical stripes and the width of the horizontal stripe between B and D _X, D _y.

Step S ₈ : 2 for each of the integral saturations S _X and S _y of each color
The integral saturations S _X and S _y between the two positions A and B are added, and the sum is
V _X and V _y . However, when D _x = 0, V _x = 0, and when D _y = 0, V _y = 0.

Step S ₉ : Obtain the average _X , _y of the integrated saturation S _X , S _y in the registers 50 _x , 50 _y corresponding to each color, and calculate the standard deviation of the integrated saturation data for the average integrated saturation from both ends of each register. For example, range (N _X / 8) + 1 excluding length 1/8
~7N _X / 8 and _{(N y / 8) + 1~7N} y / 8 in the integrating saturation data S _Xh, the S _yi calculated by the following equation.

Step S _10: the color identification matrix memory 40 by the following equation deformation the standard deviation of the pixel position relative to the maximum integration count position X _m, Y _m corresponding to each color identification matrix memory 40R~40M
For all pixels having "1" (assumed to be n) in an area excluding, for example, 1/16 of the length from the left, right, upper and lower edges,
The results are referred to as g _X and g _y . The number of sample pixels at that time is assumed to be N _X and N _y .

Step S _11: calculates the following equation.

_{Q x = V X × e X} ÷ g X ÷ n ... (20) Q y = V y × e y ÷ g y ÷ n ... (21) Step S _12: calculates the following equation.

Ux = Qx ÷ Dx ² × ｛0.5 [1 + cos (2π ・ Dx / Nx)]｝ ² … (22) Uy = Qy ÷ Dy ² × ｛0.5 [1 + cos (2π ・ Dy / Ny)]｝ ² … (23) step S _13: if wherein n is, for example n <2000, the central fringe parameter βx = Ux × {0.5 [1 -cos (π · n / 2000)]} J ... (24) βy = Uy × {0.5 [ 1−cos (π · n / 2000)]｝ ^J (25) is calculated. However, J can be set to, for example, J = 3.67.

Step S _14: In addition D _X, if a D _y Similarly K below the step S ₁₁ in the example Fig. 12 (e.g. K = 10), the weighted at the center stripe parameters β _{'X =} β _X × ｛0.5 [1-cos (2π · Dx / 2K)]｝ ^L (26) β ′ _y = β _y × {0.5 [1-cos (2π · Dy / 2K)]} ^L (27) . L is, for example, 6.6.

Step S _15: more beta _X for each color by processing, beta _y pairs or Beta'x, pairs of Beta'y, i.e. a total of 12 pieces of data are obtained. The average value Pa of the twelve data is defined as the center stripe pattern degree. The calculation unit 31 calculates the calculated center stripe pattern degree Pa
Is output.

As described above, according to the fifth aspect of the invention, it is possible to quantitatively obtain the degree of the striped pattern. Therefore, for example, the quality of the color image sensor can be determined mechanically, and a uniform product can be obtained. Also, since no manpower is required, labor saving can be achieved and cost reduction can be expected.

Although a low-pass filter having a length of 3 is used in the above-described embodiment, the length can be appropriately selected. Although a part of cos is used as the weighting function, a linear function or another function can be used.

The color image signal evaluation methods according to the first to fifth aspects of the invention described above individually evaluate color image signals by focusing on different types of color unevenness. The color image signal evaluation apparatus of the present invention described below is to comprehensively test a test object in a short time using the evaluation results for such different types of color unevenness.

FIG. 21 shows an embodiment of the total color image signal evaluation apparatus according to the sixth aspect of the present invention. In the figure, 110A, 110B, 110C, and 110D indicate different types of color unevenness evaluation devices. These color unevenness evaluation device 110A~110D are those shown in example No. 1, 6, 11, and 17 Figure, the basic degree of coloration T _s from each Hasushima pattern of I _m, peripheral stripe pattern of Pm and central stripes, and it outputs the pattern of P _a.

These evaluation output data _{T s, I m, P m} , P a multiplier 111A~1
In 11D, the coefficients E _a and E stored in the memory 112 in advance.
_b , E _c , and E _d are multiplied and given to the adder 113.

 The adder 113 calculates the sum Ω according to the following equation: Ω = Ea × Ts + Eb × Im + Ec × Pm + Ed × Pa (28)

The total sum Ω is input to the determination device 114 and compared with a reference value.
The quality is determined. The determination result is output to the display 115,
The pass / fail judgment result is displayed. Further, using the signal representing the result of the determination, for example, the non-defective product and the defective product of the color image pickup device can be automatically sorted.

The coefficients E _a , E _b , E _c , and E _d stored in the memory 112 can be obtained in advance by experiments. The combination of the coefficients can cope with, for example, a difference in the type of the color image element. Examples of the coefficients E _a , E _b , E _c , and E _d are, for example, E _a = 1.
0, E _b = 0.1, E _c = 2.0, and E _d = −1.5.

The color unevenness evaluation device is not limited to the above four types, and a color unevenness evaluation device for detecting other color unevenness may be provided.

In FIG. 21, various color unevenness evaluation devices 110A to 110D
The case corresponding to FIGS. 6, 11, and 17 has been described, respectively.However, as is apparent from FIGS. 1, 6, 11, and 17, some of the same components are included in these color unevenness evaluation devices 110A to 110D. include.

FIG. 22 is a functional block diagram of the overall color image signal evaluation apparatus configured to share these same components.

Image signal source 10, HLS conversion unit 20, hue data memory 30H,
Saturation data memory 30S, color identification matrix memory 40R-40
M, histogram calculator 31a, saturation image memory 50R-50M,
Etc. are commonly used for all or some color unevenness evaluation devices.

Evaluation of the basic coloring degree T _s according to a first aspect of the present invention after another saturation data S from the saturation data memory 30S for each predetermined block size, the average saturation in each block up to the average saturation calculation unit 34 Is calculated (step S 2 in FIG. ₂ ), the largest one among those average saturations is determined (step S ₄ ), and the maximum average saturation is determined based on the block size under the control of the control unit 100. held to the basic coloring degree registers 116A to obtain a basic coloration of T _s and cumulatively added by the adding unit 36 each time the change.

In the evaluation of the basic coloring degree T _s according to the second aspect of the present invention, the maximum average saturation calculating section 34 successively outputs the hue data H in the corresponding block of the hue data memory 30H before calculating the average saturation. For example, the luminosity correction value F corresponding to the hue data H is read out from the correction value memory 35 storing the hue luminosity characteristics as shown in FIG. 5, and the corresponding saturation is corrected by the continued correction value F. and calculating the average saturation from (Fig. 4 step S _23).

In the evaluation of the oblique pattern of I _m according to the invention of claim 3,
The correlation coefficient r was calculated by the correlation coefficient calculation unit 39 of the pixel positions in each color identification matrix memory 40R～40M "the 1" is stored (FIG. 7 step S _4), the pixel counter therewith
41 by "1" to calculate the number of pixels n stored (step S _5), by multiplying the average unit 42 from these correlation coefficient r and the pixel number n to calculate an oblique pattern of I for each color (step S _6), in this embodiment to hold the outputs their average oblique pattern of the oblique pattern of I _m diagonal pattern of register 116B.

In the evaluation of peripheral fringe pattern of P _m according to the invention of claim 4, respectively for each color identification matrix memory 40R~40M and stored contents of the saturation image memory 50R~50M by integrating count calculation unit 37 and the integrator saturation calculating unit 38 The integration counts C _X , C _y and the integration saturations S _X , S _y obtained by integrating in the Y direction and the X direction are integrated count memories 40x, 40 _y and integration saturation memories 50 _X , 50y, respectively.
To memorize.

Peripheral fringe pattern calculation section 44 determines the integral count C _X, the maximum position X _m of the C _y, the Y _m (FIG. 12 step S _3), the position X
The borders A and B of the stripe are defined on both sides of the integral saturation S _Xm and S _ym of _m and Y _m (step S ₆ ), and the integral saturation addition values V _X and V _y (step S ₈ ) of the range are defined. Integration count addition value K _X , _Ky (Step S ₉ )
Is calculated. Furthermore V _X, fringe parameters from V _y alpha _v, compute the alpha _h (Step S _10), the standard deviation d _X pixel position of "1" in the peripheral unit, to calculate a d _y (step S _12). The peripheral fringe pattern degree _Pm is calculated using these results. The result is held in the peripheral fringe pattern degree register 116c. Note that equation (7) in the sensitivity constant 0.2 and Step S ₁₀ for defining the boundary A stripe in step S _6, the B, parameter alpha _v, constants J required to calculate the alpha _h detection sensitivity due to (8) They are stored in the parameter register 43 in advance and used.

In the evaluation of the central fringe pattern of P _a according the fifth aspect of the present invention, stores the chroma saturation image memory 50R~50M from emphasizing the saturation histogram flattening part 33 (FIG. 18 step S ₄₎ . The center fringe pattern degree calculation unit 45 defines the fringe boundaries A and B in the same manner as in the evaluation of the peripheral fringe pattern degree, and calculates integrated saturation addition values V _X and V _y in the range. Further, deviations e _X and e _y of the integrated saturation at the central portion are calculated (step S ₉ ), and the standard deviation of the position of the “1” pixel excluding the peripheral portion in the color identification matrix memory 40 with respect to V _m and V _m . g _X and g _y are calculated (step S ₁₀ ). Central fringe parameters from these calculations beta _X, calculates the beta _y, and outputs an average of them as the central fringe pattern of P _a, the central fringe pattern of register 116D
To hold.

Register 116A~116D the held data _{T s, I m, P m} , P a
Are given to multipliers 111A to 111D, respectively, and the coefficient memory 11
The coefficients are multiplied by the coefficients E _a , E _b , E _c , and E _d successively from 2, and the result of the multiplication is added by the adder 113. Addition result Ω
Is compared with a predetermined reference value in the determination device 114, and if it is larger than the reference value, it is determined that the color unevenness included in the color image signal is equal to or greater than a specified value. Is determined. The determination result is displayed on the display device 115, for example.

Although the arithmetic unit 31 and the color identification unit 32 are shown in separate blocks in FIGS. 6, 11, and 17, the arithmetic unit 31 may perform the function of the color identification unit 32. That is, in the device shown in FIG.
The conversion unit 20 and all subsequent data processing can be executed by a computer. However, in order to increase the data processing speed, desired processing functions can be configured by dedicated hardware.

[Effects of the Invention] As described above, according to the present invention, a plurality of color nonuniformity evaluation devices 110A to 110D digitize and detect various color nonuniformities and standardize by multiplying the detected color nonuniformity values by coefficients. Then, the sum is obtained, and the pass / fail is determined based on whether the value of the sum Ω is higher than a predetermined value. Therefore, the color signal under test can be comprehensively inspected. Therefore, for example, by using a color image sensor for testing, it is possible to perform various color unevenness tests at once, and to perform a highly accurate inspection in a short time.

Also, by preparing coefficients to be stored in the memory 112 for each type of image sensor to be inspected, various types of color image sensors can be accurately tested.

[Brief description of the drawings]

FIG. 1 is a block diagram of an apparatus for evaluating the basic color degree according to the first aspect of the present invention, FIG. 2 is a processing flowchart for detecting the basic color degree by the apparatus of FIG. 1, and FIG. 3 is FIG. FIG. 4 is a graph showing an example of the change of the maximum average saturation with respect to the local block size in the apparatus of FIG. 4, FIG. 4 is a processing flow diagram for detecting the basic coloring degree of the invention of claim 2, and FIG. FIG. 6 is a block diagram of an apparatus for evaluating the degree of oblique pattern according to the third aspect of the present invention, and FIG. 7 is a processing flow for detecting the degree of oblique pattern by the apparatus of FIG. FIG. 8, FIG. 8 is a graph showing an example of a saturation histogram, FIG. 9 is a diagram showing a color unevenness pattern written on the color identification matrix memory when the diagonal pattern degree is large, and FIG. Color identification matrix memory when is small FIG. 11 is a block diagram of an apparatus for evaluating the peripheral fringe pattern degree according to the invention of claim 4, and FIG. 12 is a peripheral fringe pattern degree obtained by the apparatus of FIG. Flow chart for detecting the
FIG. 13 is a diagram for explaining the direction of integrating the contents of the color identification matrix memory and the X and Y integration counts obtained thereby, FIG. 14 is a graph showing an example of a weight function given to the integration count, and FIG. FIG. 16 is a graph for explaining the method for defining the integral saturation graph and the method of defining the boundaries A and B of the stripes. FIG. 16 is a diagram for explaining the operation of passing the integral saturation through a low-pass filter.
The figure is a block diagram of an apparatus for evaluating the degree of central stripe pattern according to the invention of claim 5, FIG. 18 is a processing flow diagram for detecting the degree of central stripe pattern by the apparatus of FIG. 17, and FIG. FIG. 20 is a graph showing an example of a weighting function given to the integral saturation, FIG. 21 is a block diagram of a comprehensive color unevenness evaluation device of the present invention, and FIG.
FIG. 22 is a block diagram showing another embodiment of the overall color unevenness evaluation apparatus.

Claims (6)

(57) [Claims] 1. A color image signal conversion means for converting a color image signal into saturation image data, a saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means, Repeating a predetermined number of times, averaging the chroma image data read from the chroma data memory for each pixel block of a predetermined size, and finding the maximum average chroma among them for each change in the pixel block size A basic color degree calculation unit that outputs a value corresponding to the sum of the obtained maximum average chromas as a basic color degree, and compares the basic color degree with a pass / fail judgment reference value, based on the comparison result. A color image signal evaluation device comprising: a determination unit configured to determine whether a color image signal is good or bad. 2. A color image signal conversion means for converting a color image signal into hue image data and saturation image data, and a hue data memory for storing at least one screen of hue image data from the color image signal conversion means. And a saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means; and converting the saturation image data read from the saturation data memory into the corresponding hue data. Correction is performed using the color correction value calculated based on the hue data read from the memory, and the corrected saturation image data is averaged for each pixel block of a predetermined size, and the maximum average saturation among them is calculated. Each time the pixel block size is changed, the calculation is repeated a predetermined number of times, and the value corresponding to the sum of the obtained maximum average saturations is set as the basic coloring degree. A basic coloring degree calculation unit that outputs, and compares the basic degree of coloration and quality determination reference value, the color image signal evaluating apparatus including determination means for determining quality of the color image signal based on the comparison result. 3. A color image signal converting means for converting a color image signal into hue image data and saturation image data, and a hue data memory for storing at least one screen of the hue image data from the color image signal converting means. A saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means; and a plurality of pixels each corresponding to at least one screen, provided in correspondence with a plurality of predetermined colors. A plurality of color identification matrix memories each having a storage area corresponding to a position; a color identification means for identifying to which of the plurality of colors the hue image data read from the hue data memory belongs; The saturation image data read from the memory is compared with a reference saturation, and the hue data corresponding to a pixel position having a saturation higher than the reference saturation is obtained. A color identification matrix memory writing unit that writes “1” at a corresponding pixel position of one of the color identification matrix memories corresponding to the color identified by the color identification unit in the hue image read from the color memory; Calculate the correlation coefficient for the X and Y coordinates of the pixel position where “1” is written in the color identification matrix memory,
A diagonal stripe pattern degree calculation unit that outputs a value corresponding to the correlation coefficient as a diagonal stripe pattern degree, and compares the diagonal stripe pattern degree with a pass / fail determination reference value, and determines whether the color image signal is acceptable based on the comparison result. A color image signal evaluation device comprising: 4. A color image signal conversion means for converting a color image signal into hue image data and saturation image data, and a hue data memory for storing at least one screen of the hue image data from the color image signal conversion means. A saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means; and a plurality of pixels each corresponding to at least one screen, provided in correspondence with a plurality of predetermined colors. A plurality of color identification matrix memories each having a storage area corresponding to a position; a saturation image memory provided corresponding to the plurality of colors, each having a storage area corresponding to a pixel position of at least one screen; Color identification means for identifying to which of the plurality of colors the hue image data read from the hue data memory belongs; The saturation image data read from the data memory is compared with the reference saturation, and the hue image read from the hue data memory corresponding to the pixel position having a higher saturation is identified by the color identification unit. A color identification matrix memory writing means for writing "1" at a corresponding pixel position of one of the color identification matrix memories corresponding to a color, and a color identification matrix memory corresponding to a pixel position where "1" is stored in each of the color identification matrix memories A saturation image memory writing means for writing saturation image data read from a corresponding image position of the saturation data memory into one image position of the saturation image memory to be stored; and a storage content of the color identification matrix memory At the pixel position Y
An integration count calculation unit that obtains an X integration count and a Y integration count by integrating in the direction and the X direction, respectively, and weights them by a window function determined in advance; Integrate in the Y direction and the X direction, respectively, to obtain the X integral saturation and the Y integral saturation, compare them with the reference integral saturation, and calculate the X integral saturation and the Y integral saturation exceeding the reference integral saturation. Stripe width detecting means for defining the width as the width of the vertical stripe and horizontal stripe of the corresponding color, and the sum of the saturations of the pixel positions within the stripe width detected by the stripe width detection means on each of the saturation image memories And the standard deviation weighted by the saturation of the corresponding color at the pixel location having "1" in the corresponding color identification matrix memory;
The sum of the X-integral counts and the sum of the Y-integral counts within the stripe width are calculated, respectively, and the vertical and horizontal stripes are respectively multiplied by the saturation sum, the standard deviation, and the integral count, and correspond to the multiplication results. A peripheral fringe pattern degree calculating unit that outputs a value to be performed as a peripheral fringe pattern degree, and a judging unit that compares the peripheral fringe pattern degree with a pass / fail judgment reference value and judges pass / fail of the color image signal based on the comparison result. A color image signal evaluation device comprising: 5. A color image signal converting means for converting a color image signal into hue image data and saturation image data, and a hue data memory for storing at least one screen of the hue image data from said color image signal converting means. A saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means; and a plurality of pixels each corresponding to at least one screen, provided in correspondence with a plurality of predetermined colors. A plurality of color identification matrix memories each having a storage area corresponding to a position; a saturation image memory provided corresponding to the plurality of colors, each having a storage area corresponding to a pixel position of at least one screen; Color identification means for identifying to which of the plurality of colors the hue image data read from the hue data memory belongs; The saturation image data read from the data memory is compared with the reference saturation, and the hue image read from the hue data memory corresponding to the pixel position having a higher saturation is identified by the color identification unit. A color identification matrix memory writing means for writing "1" at a corresponding pixel position of one of the color identification matrix memories corresponding to a color, and a color identification matrix memory corresponding to a pixel position where "1" is stored in each of the color identification matrix memories A saturation image memory writing means for writing saturation image data read from a corresponding image position of the saturation data memory into one image position of the saturation image memory to be stored; and a storage content of the color identification matrix memory At the pixel position Y
An integration count calculation unit that obtains an X integration count and a Y integration count by integrating in the direction and the X direction, respectively, and weights them by a window function determined in advance; Integrate in the Y direction and the X direction, respectively, to obtain the X integral saturation and the Y integral saturation, compare them with the reference integral saturation, and calculate the X integral saturation and the Y integral saturation exceeding the reference integral saturation. Stripe width detecting means for defining the width as the width of the vertical stripe and horizontal stripe of the corresponding color, and the sum of the saturations of the pixel positions within the stripe width detected by the stripe width detection means on each of the saturation image memories And the standard deviation of the saturation stored in the saturation image memory and the standard deviation of the pixel position having "1" in the corresponding color identification matrix memory are calculated in the X and Y directions, respectively. And the saturation sum and the Y direction A central stripe pattern degree calculator that divides a product of the standard deviation of saturation by the standard deviation of the position and outputs a value corresponding to each obtained result as a center stripe pattern degree; A color image signal evaluation device, comprising: a determination unit that compares the color image signal with a quality determination reference value and determines the quality of the color image signal based on the comparison result. 6. A color image signal conversion means for converting a color image signal into hue image data and saturation image data, and a hue data memory for storing at least one screen of hue image data from said color image signal conversion means. A saturation data memory for storing at least one screen of the saturation image data from the color image signal conversion means; and a plurality of pixels each corresponding to at least one screen, provided in correspondence with a plurality of predetermined colors. A plurality of color identification matrix memories each having a storage area corresponding to a position; a saturation image memory provided corresponding to the plurality of colors, each having a storage area corresponding to a pixel position of at least one screen; Color identification means for identifying to which of the plurality of colors the hue image data read from the hue data memory belongs; The saturation image data read from the data memory is averaged for each pixel block of a predetermined size, and the maximum average saturation among them is obtained each time the pixel block size is changed. A basic color saturation calculation unit that outputs a value corresponding to the sum of these maximum average chromas as a basic color saturation, and compares the saturation image data read from the saturation data memory with a reference saturation, and In the hue image read out from the hue data memory corresponding to the pixel position having saturation, "1" is assigned to a corresponding pixel position in one color identification matrix memory corresponding to the color identified by the color identification means. A color identification matrix memory writing means for writing "1", and a correlation between X and Y coordinates of a pixel position where "1" is written in the color identification matrix memory. Calculate the number,
A diagonal fringe pattern degree calculating unit that outputs a value corresponding to the correlation coefficient as a diagonal fringe pattern degree; and one chroma image corresponding to a pixel position where “1” is stored in each of the color identification matrix memories. A saturation image memory writing means for writing saturation image data read from a corresponding image position in the saturation data memory to an image position in a memory;
An integral count calculating unit that obtains an X integral count and a Y integral count by integrating in the direction and the X direction, respectively, and weights them by a predetermined window function; Integrate in the X direction to obtain the X integral saturation and the Y integral saturation, compare them with the reference integral saturation, and set the widths of the X integral saturation and the Y integral saturation exceeding the reference integral saturation respectively. A stripe width detection unit that defines the width of the vertical stripes and horizontal stripes of the corresponding color; and a sum of the saturation of pixel positions within the stripe width detected by the stripe width detection unit on each of the saturation image memories. A standard deviation weighted by the saturation of the corresponding color at the pixel position having "1" in the color identification matrix memory;
The sum of the X-integral counts and the sum of the Y-integral counts within the stripe width are calculated, respectively, and the vertical and horizontal stripes are respectively multiplied by the saturation sum, the standard deviation, and the integral count, and correspond to the multiplication results. A peripheral fringe pattern degree calculation unit that outputs a value to be performed as a peripheral fringe pattern degree; and The standard deviation of the saturation stored in the saturation image memory and the standard deviation of the pixel position having "1" in the corresponding color identification matrix memory are calculated in the X direction and the Y direction, respectively. A center stripe pattern meter that divides a product of the saturation sum and the standard deviation of the saturation by the standard deviation of the position, and outputs a value corresponding to each obtained result as a center stripe pattern degree. And a coefficient generation unit that generates a predetermined coefficient corresponding to each of the basic coloring degree, the oblique stripe pattern degree, the peripheral stripe pattern degree, and the central stripe pattern degree, and the basic coloring degree calculation unit, Multiplication means for multiplying the output from the oblique stripe pattern degree calculation section, the peripheral stripe pattern degree calculation section, the center stripe pattern degree calculation section and their corresponding coefficients, respectively, A color image signal evaluation device, comprising: an addition unit that calculates a sum; and a determination unit that compares the addition result from the addition unit with a pass / fail determination reference value and determines the pass / fail of the color image signal based on the comparison result. .