JP2807030B2 - Color image signal processing method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image signal processing method that can be used for inspecting a color imaging device made of, for example, a semiconductor.

"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 assembled as a color image sensor, color irregularities such as shading shown in FIG. 3, oblique stripes shown in FIGS. 4 and 5, and vertical stripes shown in FIG. 6 occur due to defects in the image sensor.

In other words, the screen shown in FIG. 3 should be white, and a wide area of the screen (this is indicated by reference numeral 1).
Refers to color unevenness that is colored over

The oblique stripes shown in FIGS. 4 and 5 indicate a phenomenon in which obliquely colored stripes 2 and 3 occur on a screen that should be white.

The vertical stripes shown in FIG. 5 indicate a phenomenon in which stripes 4 occur in the vertical direction.

Conventionally, a color image signal is projected on a color cathode ray tube, and this screen is visually monitored and detected.

As described above, when color unevenness caused by the color image pickup device is visually detected, it is found that the color unevenness becomes easier to see when a noise component is removed from the image signal. The reason for this is that noise contained in the imaging signal erases part of the hue image data, lightness image data, and saturation image data that indicate the presence of color unevenness, resulting in data loss at the periphery of the part where the stripe pattern should be drawn. As a result, the outline of the stripe becomes invisible, and the center and the periphery of the stripe appearing as color unevenness are hardly seen.

Therefore, when the noise component is removed from the color image data, the outline and the center of the striped pattern appearing as the color unevenness can be clearly seen, and the advantage of improving the detection accuracy of the color unevenness can be obtained.

[Problems to be Solved by the Invention] Digitized color image signals are generally converted into R, G, B monochromatic signals. In the R, G, B color space, hue, lightness, and saturation information are not independent, so when various filters such as band pass are applied to the R, G, B image data, There is a drawback that the desired image cannot be obtained due to the change.

In order to solve such a problem, the color image data is converted into hue H, lightness L, and saturation S (or may be converted into H, S, V), and image data H, L, S A method of filtering any one or all of them independently has been studied.

However, only the hue space, which represents the kind of color by angle, is the second color space.
In order to represent the same red color shown in FIG.
There is a disadvantage that it becomes discontinuous at the point of 360 degrees and 360 degrees, and cannot be digitally filtered.

The purpose in this case is to propose a color image signal processing method capable of filtering hue image data having such characteristics.

[Means for Solving the Problems] In the first invention of this application, while obtaining hue image data from a color image signal, the hue image data is projected on a complex plane, and the hue image data projected on a hydrogen plane is desired. We propose a color image signal processing method that performs filtering processing with the band characteristics of.

According to the present invention, the hue image data is projected from the discontinuity by projecting the hue image data on the complex plane, whereby the filtering process can be performed.

Therefore, it is possible to obtain the hue image data subjected to the filtering processing.

In the second invention of this application, the color image signal is converted into hue image data, lightness image data, and saturation image data, and the hue image data is subjected to a filtering process in accordance with a desired band characteristic. A color image signal processing method for obtaining an inverse tangent of a real part and an imaginary part of data, inversely converting the tangent image data to the original hue image data, and obtaining a color image signal based on the hue image data and the brightness image data and the saturation image data. Suggest.

According to the second aspect of the present invention, the reproduced color image signal has at least the hue image data subjected to the filtering process. Therefore, the hue image data is free from the influence of noise. , The outline and the center of the stripe pattern become clear, and color unevenness can be easily detected.

"Embodiment" A color image signal processing method proposed in the first invention and the second invention of this application will be described with reference to FIG.

Reference numeral 10 shown in FIG. 1 indicates a color image signal source. A red signal R, a green signal G, and a blue signal B are separately output from the color image signal source 10.

Each monochromatic signal R, G, B output from the color image signal source 10
Are converted into hue image data H, lightness image data L, and saturation image data S by a well-known HLS conversion means (or HSV conversion means) 20.

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

The hue image data H is obtained as an angle θ having a value from 0 to 360. The relationship between the color name and the angle is 0 to 30, 330 for red
~ 360, yellow 30-90, green 90-150, cyan 150-210,
Blue is 210-270 and magenta is 270-330.

Brightness is obtained in numerical values in the brightness image memory 30B, and vividness of colors is obtained in numerical values in the saturation image memory 30C.

The brightness image data L and the saturation image data S are stored in the memories 30B and 3B.
It can be sent to the inverse HLS conversion means 70 directly from 0C or through the digital filters 53 and 54 as shown, if necessary.

On the other hand, the hue image data H cannot be directly filtered by the digital filters 51 and 52 because it has discontinuous points as described above.

For this reason, in the present invention, the complex conversion means 41 for projecting the hue image data H onto the complex plane on the output side of the hue image memory 30A.
And 42 are provided.

The complex conversion means 41 indicates a portion for projecting a vector represented by hue image data hue ((i, j)) to a real part,
The part which projects the vector component represented by hue (i, j) to an imaginary part is shown.

Specifically, θ = hue (i, j) is set, the cosine and sine of θ are taken, and the complex plane realhue (i, j) and imaginaryhue (i, j)
j).

realhue (i, j) and imaginaryhue (i, j) are from -1.0 to 1.
This is an image with a real number value between 0 and stores information on Hue images.

Realhue (i, j) and imaginaryhue projected on complex plane
(I, j) is data from which discontinuities have been removed. This data is applied to digital filters 51 and 52, and REALHUE (i,
j) and IMGINARYHUE (i, j).

REALHUE obtained from digital filters 51 and 52 (i,
j) and IMGINARYHUE (i, j) are given to the inverse complex conversion means 60.

REALHUE (i, j) and IMAGINARYHUE
Put (i, j) as RTHETA = REALHUE (i, j) ITHETA = IMAGINARYHUE (i, j) and find the inverse tangent of (ITHETA / RTHETA).

The image data HUE (i, j) obtained by taking the arc tangent is the filtered hue image data.

Hue image data HUE obtained from the inverse complex conversion means 60
(I, j) is given to the inverse HLS conversion means 70 together with the brightness image data L and the saturation image data S, and the inverse HLS conversion means 70
By performing the conversion, the color image signals R, G, B can be obtained from the inverse HLS conversion means 70.

A color image can be displayed by the color image signals R, G, B.

[Effects of the Invention] As described above, according to the present invention, the filtering process can be performed on the hue image data H which could not be subjected to the filtering process because of the discontinuity. Three converted image data
H, L, S could be filtered separately respectively.

Therefore, since each of the hue, lightness, and saturation can be filtered without interfering with each other, a desired image, for example, an image in which color unevenness is emphasized can be set by arbitrarily setting a band characteristic of each filtering process. This provides an advantage that the inspection of the color imaging device and the like can be easily performed.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining a color image signal processing method of the present invention, FIG. 2 is a diagram for explaining a state of the hue space, and FIGS. 3 to 6 explain types of color unevenness. FIG. 10: color image signal source, 20: HLS conversion means, 30A: hue image memory, 30B: lightness image memory, 30C: saturation image memory, 4
1, 42: complex conversion means, 51 to 54: filter, 60: inverse complex conversion means, 70: inverse HLS conversion means.

Claims (2)

(57) [Claims] 1. A color system which obtains hue image data from a color image signal, projects the hue image data on a complex plane, and filters the hue image data projected on the complex plane with desired band characteristics. Image signal processing method. 2. A color image signal is converted into hue image data, lightness image data, and chroma image data, and the hue image data is projected onto a complex plane to convert the hue image data on the complex plane into a desired one. Filtering is performed according to the band characteristics, and the inverse tangent of the real part and the imaginary part of the filtered hue image data is obtained and inversely converted to the original hue image data, and the hue image data, the brightness image data, and the saturation image are obtained. A color image signal processing method for obtaining a color image signal from data.