JPH06118024A - Display element brightness unevenness inspection method - Google Patents

Abstract

(57) [Abstract] [Purpose] To provide an inspection method for detecting only defective luminance unevenness in a display element such as a liquid crystal display panel by distinguishing between the normal luminance unevenness caused by an illuminator and the like and the defective luminance unevenness. is there. [Structure] FIG. 1 shows an embodiment of a method for inspecting luminance unevenness of a liquid crystal display panel for realizing the present invention. An image is input from a liquid crystal display panel having a defective luminance unevenness by a TV camera, binarized, and then converted into an n × m block pattern. By comparing this block pattern with a non-defective block pattern, the uneven brightness defect is extracted. [Effect] A defective brightness unevenness of a display element such as a liquid crystal display plate is detected at high speed and in a memory-saving manner without erroneously determining a brightness unevenness inherently present by an illuminator or the like.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an output display element such as a liquid crystal display, which is suitable for inspecting the presence or absence of luminance unevenness caused by nonuniformity of luminance in one item of display failure. It is about the method.

[0002]

2. Description of the Related Art Recently, a liquid crystal display panel has replaced a cathode ray tube and a particularly small OA such as a personal computer or a word processor.
It has been widely used in equipment. Moreover, not only for small-sized liquid crystal display panels, these liquid crystal display panels are in the trend of large screen and high definition. In the display quality inspection item in the inspection department in the manufacturing process of this liquid crystal display panel, one dot or several dots of the liquid crystal display element becomes a defect or a bright spot.
There are "line omission defect", "brightness unevenness defect" which causes non-uniformity of brightness in a wide area of the display surface, and the like. Conventionally, these items have been mainly inspected by visual sensory evaluation, but in recent years, automatic inspection has been proposed.
Regarding dot dropouts and line dropouts, there are lighting inspections and inspection methods using a reference pattern, as in JP-A 1-189696 and JP-A 1-191048. The problem with automation of this point defect inspection is the inspection processing time, and a processing speed equivalent to visual inspection is required. Further, the luminance unevenness evaluation has many subjective factors in the evaluation, which makes it difficult to quantify and automate the evaluation. As an example of this automatic evaluation, Japanese Patent Application Laid-Open No. 2-193271 has been filed. This is to measure the brightness unevenness of the TFT, input the image with the TV camera, and then (original image)-(shifted image), (original image)-
By performing the processing (images with time shift), it is intended to evaluate the occurrence state of unevenness from the luminance dispersion of each difference image.

[0003]

A display element such as a liquid crystal display panel has a built-in illumination such as a fluorescent lamp. On the display surface, the place where the illuminator is attached becomes bright, and a luminance difference from the peripheral portion occurs, causing uneven luminance. Therefore, the brightness unevenness can be seen even in a normal good product. On the other hand, there is brightness unevenness that occurs due to non-uniformity of the liquid crystal thickness, and this is defective brightness unevenness that must be detected in the display element inspection process. In the conventional method, the determination as to the luminance unevenness of the non-defective product is not taken into consideration, and all of them have been evaluated as defective luminance unevenness. An object of the present invention is to provide an inspection method capable of distinguishing and detecting the brightness unevenness peculiar to the non-defective product and the brightness unevenness defective.

[0004]

According to the present invention, luminance unevenness is detected by using image processing in order to make an evaluation close to visual inspection.
As a general image processing method, a method of storing a non-defective luminance unevenness image, obtaining a difference image between it and the image of the display element to be inspected, and leaving the remaining image as defective luminance unevenness is possible, An image of 512 × 512 pixels requires a long processing time, and a huge memory capacity is required to store a non-defective image for each model of display element. Therefore, the inventors have invented the luminance unevenness inspection method of the present invention which solves these problems. To do so, first, the input image is binarized, and the result is converted into an n × m block pattern, which is coarser than the original image.
Then, the block patterns of the same size are stored in advance for non-defective products, and the uneven brightness portion is extracted by comparing with the image to be inspected.

[0005]

By converting the image into a block pattern in this way, the amount of data is compressed to enable high-speed processing and memory saving, and by comparing it with a non-defective pattern, non-defective non-uniform luminance is selected. Can be detected separately.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to FIGS. 1 to 6 by taking an example of inspection of brightness unevenness in a liquid crystal display plate such as a TFT or LCD. When a liquid crystal display panel is imaged by a TV camera, uneven brightness occurs on the panel as shown in FIG. At this time, the TV image size is 512 × 512 pixels, 2
56 gradations. The processing procedure for detecting this brightness unevenness is shown in the flowchart of FIG. First, the luminance unevenness in FIG. 1A is binarized. For the binarization, a threshold is set so that the shape can be extracted well. For example, as an example of the means for determining the optimum binarization threshold value for extracting the brightness unevenness shape of the liquid crystal display panel of the present embodiment, 1/2 or 1/4 of the total area in the binarization result, 1 There is a method in which the brightness / luminance value at which / 8 becomes $ FF (bright part) is used as the threshold value. Figure 1 (2)
Is the binarization result when the threshold value is 1/4. The method of determining this threshold differs depending on the inspection target,
Whichever object is used, 2
A method of selecting a threshold for quantification is used.

Next, the binarized image is converted into an n × m block pattern to compress the data. In the example of the processed image of FIG. 1B, the image of 512 × 512 pixels is converted into the 6 × 8 block indicated by the broken line. As an actual process, if 1/2 or more is $ FF in the range of 85 × 64 pixels which is one block in FIG. 1 (2), that block is $ F.
Set to F. As a result of processing this on the entire surface of FIG. 1 (2), 12 blocks with a convex right side are $ F as shown in FIG. 1 (3).
It becomes a block pattern of F. The above-described processing is performed on each liquid crystal display panel to be inspected, and each block pattern is obtained. In this example, the blocks are 6 × 8, but if there are many blocks, this block pattern will be the same as the original image, and if it is too small, the characteristic shape of uneven brightness will collapse. The number of blocks is set to a number at which features can be easily obtained according to the difference in the inspection target and the size of the uneven brightness.

Next, as shown by the broken line in the flow chart of FIG. 4, it is determined whether or not there is a luminance non-uniformity defect from the previously obtained block pattern and the non-defective block pattern. Therefore, a good block pattern is stored in advance. FIG. 2 shows an example of the block pattern of the non-defective product to be stored.
(A) and (b). Even in such a non-defective block pattern, the place where the illuminator of the liquid crystal display panel is located has particularly high brightness as compared with the surrounding area, and brightness unevenness constantly occurs. For example, (a) is a type in which the light 2 for backlight illumination is provided only on the left side of the liquid crystal display panel. The block pattern at this time is a pattern in which the center of the convex shape to the right is slightly bright as shown in the right figure of (a). Further, (b) is a type in which the fluorescent lamps 2 are arranged in two horizontal rows in the center of the liquid crystal display plate, and the block pattern has two horizontal lines. As described above, even a normal good product has a brightness unevenness pattern depending on the illumination state, and a good product block pattern needs to be obtained and stored for each illumination type.

As described above, the quality non-uniformity block pattern of the non-defective product, which is obtained in advance, is compared with the evaluated pattern to determine the quality of the inspected liquid crystal display panel. There are various methods for this determination, and one example thereof is (1) a non-defective pattern as shown in FIG. 3, and (2) an evaluated pattern having a luminance unevenness defect in the right corner of the screen is subtracted. As a result, Fig. 3 (3)
It is possible to extract uneven brightness, which is a mismatch point. In the process of the flowchart of FIG. 4, if there is such a disagreement point, it is determined that the luminance unevenness is defective, and if there is no disagreement point, it is determined as a good product.

As another method for determining the presence / absence of brightness unevenness, brightness unevenness due to illumination is targeted vertically or horizontally, and on the contrary, brightness unevenness appears irregularly. There is a method of determining that there is uneven brightness. Further, in the case of a display element in which there is no unevenness in brightness due to lighting or the like, there is a method of setting a threshold value for the number of blocks of $ FF in the block pattern and determining the presence / absence.

An example in which the brightness unevenness evaluation method of the present invention is mounted on a microcomputer and carried out on a liquid crystal display panel is a brightness unevenness inspection apparatus shown in FIG. The liquid crystal display panel 1 to be inspected is XY on the surface plate 5.
It is fixed on the table 4 and can move freely. The liquid crystal display panel 1 is brought into a display state, and the state at that time is imaged by the TV camera 3 from above. The image is input to the microcomputer 6 in the surface plate 5, and the uneven brightness detection processing is performed. The non-defective pattern for comparison is stored in the microcomputer 6 in advance. In the microcomputer 6, a method of processing all the above-mentioned brightness unevenness detection steps by software, a binarization processing circuit having an optimum binarization algorithm, a block patterning circuit,
There is a method of inspecting using hardware such as an image comparison circuit.

Although the brightness unevenness of the liquid crystal display plate has been described above as an example, this method can also be applied to the evaluation of brightness unevenness and color unevenness of a CRT display.

Further, in this example, 2 is set before the block patterning.
Although the binarization process is performed, a multi-valued image that is not only binarized but also ternarized and quaternized may be used. It becomes a contour line, and the processing of the block patterning becomes a little complicated, but the shape of the uneven brightness can be expressed more accurately.

[0014]

By performing the above processing, it is possible to detect a defective luminance unevenness of a display element such as a liquid crystal display plate without erroneously recognizing it as a luminance unevenness inherently present by the illuminator or the like. Also, by converting an image of 512 × 512 pixels into a 6 × 8 block pattern, the memory capacity and processing time can be shortened to 1/5461.

[Brief description of drawings]

FIG. 1 is a diagram illustrating an example of a processing procedure of a liquid crystal display panel luminance unevenness inspection method.

2A and 2B are diagrams showing a normal brightness unevenness block pattern of a liquid crystal display panel, FIG. 2A is a type having an illuminating fluorescent lamp on the left, and FIG. 2B is a diagram having an illuminating fluorescent lamp in the center.

FIG. 3 is a processing diagram for determining the presence / absence of luminance unevenness on a liquid crystal display panel.

FIG. 4 is a flowchart of a liquid crystal display panel luminance unevenness inspection method processing.

FIG. 5 is an external view of a liquid crystal display panel luminance unevenness inspection device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Liquid crystal display board, 2 ... Illuminator, 3 ... TV camera, 4 ... XY table, 5 ... Surface plate, 6 ... Image processing apparatus.

Claims (1)

[Claims] 1. In a display inspection of an output element such as a liquid crystal display panel, non-uniformity in brightness caused by non-uniformity in brightness is converted into an n × m block pattern after binarization, and a non-defective pattern similarly formed into a block pattern. A method for inspecting luminance unevenness of a display element, which comprises detecting unevenness in luminance by comparison with.