JPS63200278A - Image quality inspection apparatus - Google Patents

Abstract

PURPOSE:To detect the variance in density contained in an ambiguous video signal in accordant with human visual characteristics, by dividing a memory area into smaller pieces and performing comparison of the evenness of image quality among those divided small areas. CONSTITUTION:A light box 1 radiates white light on an image pickup device 2 to be inspected. Video signals given from the device 2 are stored in an image memory 4 of an image quality inspector 10 via a decoder 3. A density variance detection controlling part 5 divides the memory area of the memory 4 into smaller pieces and obtains the evenness in picture quality for each of these divided small areas. In case the evenness of a noticed small area is higher than those of its nearby small areas, said small area is extracted as the maximum small area. Then this small area is decided as a density variance area if the evenness obtained by normalizing once the evenness of said maximum small area is higher than a prescribed threshold value.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Field of Application) The present invention is directed to the prevention and control of defects such as stains that exist in video signals from transportation and equipment. This invention relates to an image quality inspection device that inspects for one-degree unevenness.

(Prior art) For example, as a CCD color image sensor, red,
It is known that three primary color filters of blue and green are attached to each CCD element and the image is divided into the three primary colors and read. However, if the attachment of the filter is uneven or incomplete, color unevenness <density unevenness) will occur in the obtained video signal, making it impossible to obtain a high-quality image.

As a device for detecting the depth unevenness that fundamentally exists in a video signal as described above, a device using a differential binarization method is conventionally known. This device filters a video signal containing uneven turbidity using a differential filter, then binarizes it using a preset threshold to detect the boundary between the uneven turbidity and the background.

However, although the conventional device described above can easily detect the boundary portion when the difference between the density unevenness and the background is clear, the density changes slowly and the boundary portion between the 11° unevenness and the background is unclear. In clear cases, sufficient detection is not possible.

This will be explained in more detail using FIGS. 6 and 7. FIG. 6 shows each pixel row of a video signal with uneven density, and FIG. 7 shows each pixel row of a normal video signal.

6(A) and FIG. 7(A) are pixel arrays having a 16×16 configuration, and the numbers in the figures indicate the density value of each pixel. Furthermore, it indicates the 6th S degree value. The boxed area in FIG. 6 is an uneven density area.

Comparing Figure 6 (B) and Figure 7 (13), the boxed area in Figure 6 ([3) is an area with uneven turbidity, but the difference from the dorsal and neck area is small; It is difficult to distinguish between the back part and the uneven turbidity part. Also, even if we compare it with the normal image signal in Figure 7 ([3)], we cannot tell where the turbidity is uneven.

On the other hand, in the conventional example described above, the differential filter processing is based on the absolute difference between the center pixel of the filter and the neighboring pixels. 1Iil with a density difference of ±20 in a video signal with a height of 1i4111 degrees with hTi of 100 where there is density unevenness with a 1α difference of ±10! When there are q degrees of unevenness, the result after differential filtering is more appealing. However, the difference in density of the former, which has a relative difference of 111 degrees compared to the human visual characteristic t, is strongly felt.

Furthermore, there is a characteristic that the presence of only one smear is felt more strongly than when there is uniformly smudge-like turbidity unevenness throughout the image.

As described above, the conventional method has a problem in that turbidity unevenness is not detected in a way that matches the visual characteristics of humans.

<Problems to be solved by the invention> 1) As mentioned above, with conventional image quality inspection equipment, it is difficult to inspect when the difference between the uneven turbidity area and the backwater area is unclear. Ta. In addition, since no test was conducted based on human visual characteristics, the test results were insufficient.

The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to be able to detect uneven turbidity in a video signal in which the difference between the uneven turbidity portion and the background portion is unclear. An object of the present invention is to provide an image quality inspection device capable of detecting 81-degree unevenness that is compatible with human visual characteristics.

[Structure of the Invention] <Means for Solving the Problems> In order to solve the problems mentioned in -1-, the present invention inspects the presence or absence of density unevenness contained in a video signal of an object to be inspected, such as an imaging device. The apparatus comprises: storage means for converting video signals from the test object, which is a decoy image of a reference white light source, into one and storing it in X×Y areas; and the storage area on the image storage means. a dividing means for dividing the image into m×n small regions, a uniformity performance means for determining the uniformity of image quality in each small region fu, and a uniformity ratio between the rough small region and neighboring small regions to which the small region is adjacent. A local maximum/small area extraction means that compares the uniformity of the small area and extracts it as a local maximum/small area if the uniformity of the small area is greater than the uniformity of any of the neighboring small areas, and a local maximum/small area extraction means that normalizes the uniformity of the extracted local maximum/small area. and determining means for determining the small area as a turbidity uneven area when the normalized average power is equal to or higher than a preset threshold value. 1. (Operation) The present invention In order to test whether turbidity unevenness is included in the image signal of the silver imager, first, a reference white light source is layered, and the image signal is stored in one stage of memory-T.

Next, the storage area is divided into mxn small areas,
The uniformity of image quality is determined for each small region.

Then, the uniformity of the focused small area is 11g1
If the uniformity is greater than any of the adjacent small regions, the small region is extracted as a local maximum and small region, and the uniformity after normalizing the uniformity (maximum value) of this extracted local maximum and small region is determined in advance. If the value is greater than or equal to the set threshold, the small area is determined to be an uneven density area.

(Embodiment) FIG. 1 shows the configuration of an embodiment of the apparatus according to the present invention.

The image quality inspection device 10 of this embodiment takes in the video signal S1 of the image carrier 2, which is an image of the illumination light of the light box 1, and examines the density unevenness such as spots contained in this visualized pJS1 by human visual characteristics. It is something that should be considered and detected.

The light box 1 has a white light source serving as a reference light source,
White light is irradiated toward the imaging device 2 to be inspected.

The image carrier 2 is, for example, a COD color image sensor 1, and each pixel has i, f5, . It consists of a green color filter attached. Previous) As I did 23,
Due to non-uniformity, imperfection, etc. when these color filters are attached to each element, there is a possibility that the video signal of the image carrier 2 fundamentally contains density unevenness in the form of spots.

The image quality inspection device 10 includes a decoder 3 that zumbling and quantizing the video signal S1 from the image transfer device 2, and a decoder 3 that zumbling and quantizing the video signal S1 from the image transfer device 2,
In order to control the entire apparatus and detect unevenness in density from the video signal stored in the image memory 4, the image memory 4 stores data in a size of 256 pixels, and can execute the process 4 shown in FIG. Ia unevenness detection control section 5 and video signal S1
The display unit 6 includes a CRT display that displays images and displays the position of detected density unevenness.

The present embodiment has the above-described configuration, and its operation will be explained below with reference to FIG. 2 and subsequent drawings.

As shown in the broach 1r-t of Fig. 2, a video signal S based on the reference white light imaged by the photographing device 2 to be inspected.
1 is quantized by the decoder 3 and stored in the image memory 4. (Step 201). The size of the image is X×Y=256×256 (picture!#) as shown in FIG.

Next, as shown in Fig. 3, the input data is divided into 31 x 16 small areas such that each small area overlaps by half (eight pixels) by 12J in the X direction, in a 3 x 16 configuration. , each small area has m=l~31 in the X direction, and n=1 in the Y direction.
It is divided into 1 to 16, and its position is indicated by (m, mouth).

Next, a process for determining the density uniformity of each divided small area is executed (steps 203 to 215).
The uniformity of 11m is the standard deviation of 11a of 3mn (m = 1
~31. n=1 to 16).

First, in step 203, initialization processing is performed and standard deviation Smn=0. It is assumed that y = i and n = i. Then, a deviation of 3 mn from the small region of (m, n)=(1, 1>) is determined from the sequentiality.

First, the average temperature P within a small area is determined by the following equation (1).

Also, its standard deviation Smn is is required.

Above) Like a flea, once the standard deviation S11 of the small region (1, 1) is found, next, the small head 1i! The standard deviation S2+ of (2, 1) is found. In this way, the small frf region (
31, 16), each standard deviation S I
n is determined (steps 205° to 215).

In the next steps 217 to 233, 1
A process for determining whether there is an 11 degree unevenness is executed.

Zunawara, small area (2, 2> to small head area (30°15)
The uniformity of each small region of paralysis with its eight neighboring regions is compared. For example, the small area (2,
Regarding the uniformity 822 of 2>, its 8 neighboring regions are (1.1) and (1,2).

(1,3), (2,1), <2. 3),
(3,1), (3,2> and (3,3) uniformity S tube 1°S+ 2 , 82 + , 823 , 83
+, 832 and S33 are compared. As a result, if the uniformity 822 is larger than any of the eight neighboring uniformities, and if the uniformity 822 is the maximum value, a small region (2, 2> is extracted as a maximum/small region).

Then, this uniformity (maximum value) S22 is normalized by the following general formula in order to obtain a relative ratio with the neighboring uniformity.

However, (i, j)=(m, n).

Next, the magnitude of the uniformity S=mri after normalization and a preset threshold value lb is determined, and this uniformity S=mri
If n is greater than or equal to the threshold Th, the small area is determined to be “lHI [unevenness”].

As the value of the threshold Th, a value of 2 (2) is appropriately selected in consideration of human visual characteristics.Then, the position of the density unevenness is displayed on the display unit 6.

Next, the operation of this embodiment will be explained in detail.

Figure 4 (A> shows the uniformity 5Iln of each small area of a video signal containing temperature unevenness, and Figure 5 (A) shows a good video signal, respectively. ) and FIG. 5(B) are obtained by calculating the maximum value of uniformity 3nn in FIG. 4([3) and FIG. The results are shown.The solid line in Fig. 4 indicates the density unevenness region.

Comparing the uniformity of the density unevenness area shown in FIG. 4 (Δ) and the broken line area in FIG.

However, when comparing Figure 4 (B) and Figure 5 N3), after normalizing the maximum value, the temperature unevenness Wlf has a value of 18
It can be seen that the broken line area of the non-defective sample with 1 is 127, and the value of the uneven density area is larger.

Therefore, for example, if the threshold value is set to a value of 150, the area surrounded by the solid line in FIG. 4(B) will be detected as temperature unevenness, but
The area surrounded by the broken line in FIG. 5(B) is not considered to be temperature unevenness.

In this way, in this embodiment, compared to the conventional examples shown in FIGS. 6 and 7, it is possible to detect density unevenness in a video signal where the boundary between the uneven portion and the background portion is unclear. . Also,
Since the determination of 11 degree unevenness is made based on the relative uniformity with neighboring areas rather than the absolute uniformity, it is possible to accurately detect image unevenness in consideration of human visual characteristics.

[Effects of the Invention] As explained above, according to the present invention, it is possible to detect temperature unevenness in a video signal in which the difference between the temperature uneven part and the background part is unclear.
Furthermore, it is possible to accurately detect ii unevenness that is compatible with human visual characteristics.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the apparatus according to the present invention, FIG. 2 is a flowchart showing the processing procedure of an embodiment of the apparatus according to the present invention, and FIG. 3 is a configuration diagram of an image memory. 4 and 5 are explanatory diagrams of the operation of an embodiment according to the present invention, and FIGS. 6 and 7 are explanatory diagrams of the operation of a conventional example. 1... Light box 2... IfJi image device 3... Decoder 4... Image memory 5... Temperature unevenness detection control section 6... Display section 10... Image quality inspection device

Claims (2)

[Claims] (1) A device that inspects the presence or absence of density unevenness contained in a video signal of an inspection target such as an imaging device, which quantizes the video signal from the inspection target that captures an image of a reference white light source and performs X×Y a storage means for storing the image into three areas; a dividing means for dividing the storage area on the image storage means into m×n small areas; a uniformity calculation means for calculating uniformity of image quality for each small area; Compare the uniformity of the focused small region and neighboring small regions to which the small region is adjacent, and if the uniformity of the small region is greater than the uniformity of any of the neighboring small regions, extract it as a local maximum/small region. a local maximum/small area extracting unit; and a determining unit that normalizes the uniformity of the extracted local maximum/small area and determines the small area to be an uneven density area if the normalized uniformity is equal to or higher than a preset threshold. An image quality inspection device comprising: (2) The image quality inspection apparatus according to claim 1, wherein the dividing means is a means for dividing the divided small areas so that they overlap with each other.