JP2003167530A - Method and device for display picture inspection - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a display picture inspecting method for accurately and precisely inspecting a display. <P>SOLUTION: In the method of inspecting respective cells in a specified display area on a display screen by measuring the positions of the respective cells constituting individual pixels by using an image which is picked up by using an image pickup means on the display screen where a position measurement pattern is displayed, providing luminance measurement areas at the measured positions of the respective cells by using an image which is picked up by the image pickup means on the display screen where an inspection pattern is displayed, taking measurements of luminance in the respective luminance measurement areas, computing the mean value or center value of obtained luminance measurement results, and comparing the value with a threshold, when the luminance mean value or center value of the respective cells is compared with the threshold, and when a cell with high luminance is present nearby a cell to be inspected, the threshold is made larger according to the luminance values of cells nearby the cell. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a display screen inspection method and a display screen inspection device. More specifically, the present invention relates to a display screen inspection method and a display screen inspection apparatus for inspecting a display screen represented by a flat panel display such as a liquid crystal display.

[0002]

14-16 show, for example, Japanese Patent Application Laid-Open No. 6-2.
It is a figure which shows the inspection method of the conventional liquid crystal display panel shown by 50139 gazette. First, the inspection device will be described. In FIG. 14, the inspection table 51 has Xs orthogonal to each other,
It is slidable in the biaxial direction of Y and rotatable in the horizontal direction, and the liquid crystal display panel 55 is mounted on the upper surface thereof. The inspection table 51 is provided with an opening 51a at a portion corresponding to the liquid crystal display area. A backlight 52 is arranged below the inspection table 51, and the light emitted from the backlight 52 is emitted from the opening 5.
The liquid crystal display panel 55 is illuminated through 1a. The unloader 5 for removing the liquid crystal display panel 55 for which the measurement is completed from the inspection table 51 on the left and right of the inspection table 51.
3 and a loader 54 for mounting a liquid crystal display panel 55 for new measurement on the inspection table 51. A liquid crystal display panel 55 is provided on the side of the inspection table 51.
A prober 56 for illuminating the liquid crystal display panel 55 is arranged by connecting to the connection terminal of. A CCD camera 57 in which photocells are arranged in a matrix is arranged above the liquid crystal display panel 55 and takes in the lighting brightness of the liquid crystal display panel 55. An image processing device 58 having an AD or AC converter 59, an arithmetic circuit 60, an image memory 61 and a display circuit 62 is connected to the CCD camera 57.

In the inspection method using such an inspection device, the liquid crystal display panel 55 is imaged by the CCD camera 57,
The inspection is performed by providing a brightness measurement area for each cell that constitutes each pixel and obtaining the average value of the brightness in the area. FIG. 15 is an image captured by the CCD camera 57, which includes R (red), R (R1 (display pixel in the first row)), R
(R2 (display pixel in the second row)), G (green), and B (blue) represent the position of each cell constituting RGB of one pixel of the liquid crystal,
The photocells S arranged in a matrix indicate one light receiving element of the CCD camera.

Next, the operation will be described. In this conventional inspection method, the CCD camera is first aligned. In alignment, for example, as shown in FIG. 16, position measurement on an image is performed at predetermined three points (alignment display pixels M1 to M3) on the screen of the liquid crystal display panel, and the distance on the image and the actual position are measured. The reference distances Xo and Yo in the XY directions are compared. Then, the position of each cell is specified by calculating the position and inclination of the liquid crystal display panel. In FIG. 16, reference numeral 65 denotes the entire display area,
.About.65C is a visual field section. Next, a luminance measurement region is provided based on the position of each cell, and the average luminance value in the region is compared with a predetermined threshold value to inspect each cell.

[0005]

However, in the conventional inspection method, since the average luminance value and the predetermined threshold value are simply compared, an accurate inspection cannot be performed. Because
When there is a bright spot defect cell with high brightness (a cell that cannot be controlled and is always lit), the brightness of the defective cell causes light to circulate and the measured brightness of the surrounding cells will rise. Is caused. Since the influence of brightness extends around the two cells in the surroundings, when compared with a predetermined threshold value, the peripheral cells of the bright spot defective cell are erroneously recognized as a defect, and there is a problem that the correct inspection cannot be performed.

Further, in the conventional alignment method of the inspection method, the alignment is executed simply by using the measurement results of the predetermined positions on the display screen and the actual distances. Since there is lens distortion of the CCD camera, it is difficult to accurately measure the position of the cell. Further, if a slight error occurs in the position measurement of the CCD camera due to elastic deformation of the mechanism for holding the CCD camera or backlash, there is a problem in that the alignment is misaligned. In the case of liquid crystal, since the short side of one cell is 100 μm or less, a slight misalignment is not allowed.

Further, in the conventional inspection method, although the details of the image pickup method are not described, for example, the dynamic range of the brightness of the liquid crystal display panel is very large, and the dynamic range is about 100 times that of the existing ordinary CCD camera. is necessary. Therefore, in order to measure and inspect from low-luminance cells to high-luminance cells, there is a problem that imaging under normal fixed conditions is extremely impossible.

Further, in the conventional inspection method, the details of the measuring method are not described, but regarding the cells of different colors, the sensitivities of colors differ depending on the image pickup means, and
Since the transmitted light quantity characteristic of the color filter differs depending on the color, there is a problem that accurate inspection cannot be performed if the same threshold value is used in each cell having different colors of liquid crystal.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a display screen inspection method and a display screen inspection apparatus capable of accurately and accurately inspecting a display. To do.

[0010]

A method for inspecting a display screen according to claim 1, wherein an image of a display screen on which a position measurement pattern is displayed, which is imaged by an imaging means, is used for a predetermined display area of the display screen. Then, the position of each cell that constitutes each pixel is measured, and the brightness is measured at each of the measured positions of the cells using the image of the display screen on which the inspection pattern is displayed, which is imaged by using the imaging unit. A method is provided in which an area is provided, the brightness is measured in each brightness measurement area, the average value or the median value of the obtained brightness measurement results is calculated, and the result is compared with a threshold value to inspect each cell. , When comparing the brightness average value or median value of each cell with the threshold value, if there is a cell with high brightness in the vicinity of the cell to be inspected, depending on the brightness value of the cells in the vicinity. Characterized by increasing the threshold.

According to a second aspect of the present invention, in the display screen inspection method, the luminance measurement is performed a plurality of times in each of the luminance measurement areas by changing the exposure time of the image pickup means, and the obtained luminance measurement result is obtained. By synthesizing, the luminance in each luminance measurement area is calculated.

According to a third aspect of the present invention, there is provided a method for inspecting a display screen, wherein a specific coordinate position on the display screen and a position on the image obtained by the image pickup means relating to the coordinate position are stored in advance.
The absolute position of the image pickup unit and the conversion formula on the display screen are calculated by using a plurality of sets of data of the absolute position of the image pickup unit at that time, and the absolute position of the image pickup unit and the conversion formula are calculated at the time of inspection. Measure the expected position on the image of the measurement pattern obtained by using, and measure the deviation of the position on the image of the actual measurement pattern, and use this deviation amount to newly calculate the conversion formula for the image coordinates and the display coordinates, Using this conversion formula, the position of the inspected cell on the image is converted into the display coordinate position and output.

According to a fourth aspect of the present invention, in the display screen inspecting method, a plurality of cells at predetermined positions on the display screen are turned on in the position measuring pattern, the positions of a plurality of cells are measured, and they are set in a small area. By dividing and performing a curve regression of the relationship between the image coordinates and the display coordinates for each of the small areas, the amount of distortion of the image of the image pickup unit is corrected,
Calculate the position of each cell.

According to a fifth aspect of the present invention, a display screen inspection method calculates an average value or a median value for each cell of different colors in the entire inspection area for creating the threshold value,
A threshold value is created for each cell having a different color by adding a reference value to the average value or the median value.

According to a sixth aspect of the present invention, in the display screen inspection method, the luminance measurement is corrected for each cell having a different color by multiplying the result of the luminance measurement by a coefficient based on the sensitivity characteristic of the image pickup means.

Further, the display screen inspection apparatus according to claim 7 is an image pickup means for picking up an image of the display screen, a display screen control means for controlling the screen of the display screen, and a moving means for holding and moving the image pickup means. And a measuring means for measuring the absolute position of the image pickup means, and an image of the display screen on which the position measurement pattern is displayed, which is picked up by the image pickup means, is processed to measure the position of each cell forming each pixel. Then, the image of the display screen on which the inspection pattern is displayed, which is picked up by the image pickup means, is processed, and the brightness measurement regions are provided at the positions of the respective measured cells, and the brightness measurement is performed in each of the brightness measurement regions. The average value or median value of the obtained brightness measurement results is calculated, and compared with the threshold value, each cell is inspected, and the brightness average of each cell is calculated. Or, when comparing the median value and the threshold value, if there is a cell with high brightness in the vicinity of the cell to be inspected, the threshold value is increased according to the brightness value of the cells in the vicinity. The image processing means and the control means for controlling the entire flow of the inspection are provided.

[0017]

First Embodiment FIG. 1 is a diagram for explaining a display screen inspection method and a display screen inspection apparatus according to a first embodiment of the present invention. In FIG. 1, 1 is an image pickup means such as a CCD camera, 2 is a display screen control means, 3 is a moving means for moving the image pickup means 1, 4 is an image pickup position measuring means, 5 is an image processing means, 6 is a control means, 7 is an output value of the determination result. Reference numeral 8 denotes a display device represented by a liquid crystal display panel which is an inspection object.

Next, the operation will be described. First, a rough inspection is performed on the entire display screen, and the image pickup means 1 is moved to an approximate location where a defect exists. The rough inspection may be performed directly by a person, or may be performed by another device and the image pickup means 1 may be moved based on the result.
When a person performs a rough inspection, the display screen control means 2 controls the display device 8 to display a screen pattern suitable for the rough inspection, such as an all black screen. The movement of the image pickup means 1 is performed by the movement means 3 and is movable in the XY directions. This operation can be automatically performed based on the signal of the control means 6, but may be manually moved by a person. In the first embodiment, the moving means 3 holds the image pickup means 1 and is located above the display device 8. However, the image pickup means 1 is fixed and another moving means holds the display device 8. It can also be configured to move. This is because the function of moving the visual field of the image pickup means 1 is equivalent. Then, after a screen control instruction is sent from the control means 6 to the display screen control means 2, the measurement pattern is displayed on the display device 8. The image capturing means 1 captures an image of part or all of the display device 8 on which the measurement pattern is displayed, and stores the image in the image processing means 5. Then, after the control means 6 sends a screen control instruction to the display screen control means 2, the inspection pattern is displayed on the display device 8. The image pickup means 1 is a display device 8 on which an inspection pattern is displayed.
A part or all of the image is captured and the image is stored in the image processing means 5. At this time, two or more types of patterns may be applied to the measurement pattern and the inspection pattern by changing the exposure time. Further, the display order of the measurement pattern and the inspection pattern does not necessarily have to be the measurement pattern first, and the inspection pattern can be displayed first. In general, it is preferable to generate a pattern so as to have the shortest time in consideration of the time required to generate the pattern and the switching time between patterns. Further, when the inspection pattern is imaged, the exposure time of the image pickup means 1 is set to, for example, 1/2 s, 1 /
For example, 20s, 1 / 200s, etc. are switched a plurality of times to capture an image, and the image is accumulated. Depending on the special image pickup means, the same processing may be performed in the internal circuit of the image pickup means, but in that case, that function may be used.

FIG. 2 is a flow chart showing the inspection procedure for the accumulated image. In FIG. 2, in step S1, the image of the measurement pattern is image-processed, and the position of each cell in the image on the image is measured. Step S
In 2, based on the position on the image of each cell obtained in step S1, the luminance measurement region is provided at the light emission position of each cell from the image of the inspection pattern to perform the luminance measurement, and the average value of the luminance of each cell or Measure the median. In step S3, it is checked whether or not there is a high-intensity cell around the cell to be inspected, and if it exists, the threshold value is changed (step S).
4). Then, in steps S5 to S7, the threshold value is compared with the average value or the median value of the luminances of the cells to be inspected to judge the quality of the cells of the display device. These steps S3 to S7 are performed for all the cells or a part of the cells shown in the image of the inspection pattern, and then in step S8, they are repeated until the determination of all the cells is completed. After the determination of all cells is completed, the image position of the defective cell is converted into display screen coordinates in step S9, and the presence or absence of the defective cell, the brightness information of the defective cell, and the defect position are output in step S10.

Next, the contents of each step will be described in detail.

In step S1, the image of the measurement pattern is image-processed and the position of each cell in the image on the image is calculated. 3 to 4 show a cross hatch image CP as an example of the measurement pattern. Although the reason will be described later, when the display device is a color TFT liquid crystal display device, the crosshatch is configured using only cells of a specific color.
In FIGS. 3-4, the crosshatch lines appear not to be continuous, especially in the X direction, but this is because, for example, only the G (green) cell is illuminated, and the R (red) and B cells.
This is because the (blue) cell is turned off.

The image processing means 5 first measures the positions of a plurality of cells at predetermined positions by using the image of the measurement pattern. FIG. 5 shows the center positions P of the plurality of cells measured from the measurement pattern. FIG. 6 shows an example of a measurement method for performing the measurement of FIG.
The projection in the XY directions, which is a known technique, is used. Figure 6
As shown in (a) and (b), since the approximate positions of the cross hatch lines L1 and L2 can be calculated by projection,
A measurement area Z is provided at that position, and as shown in FIG. 6C, the exact position of each vertical and horizontal line LY, LX is calculated by the calculation of the center of gravity in the area, and the intersection is defined as the center of the predetermined cell C. Position P. At this time, in calculating the center of gravity, it is desirable that the crosshatch line is composed of only specific colors. Since the brightness of cells with different colors differs depending on the sensitivity of the image pickup means and the transmission characteristics of the color filters that compose the liquid crystal panel, when the line is composed of cells of different colors, the center of gravity position is significantly displaced from the center position of the cell. Because there is. As shown in FIG. 5, the center position P of the plurality of cells C
After calculating these, the center position P and basic data of the known crosshatch (color of cells forming the crosshatch, how many cells are arranged in the crosshatch interval, vertical and horizontal lengths of the cells) The exact position on the image of all cells displayed on the screen is obtained by combining and calculating.

In step S2, the image processing of the image of the inspection pattern is performed as follows. The flowchart of step S2 is shown in FIG. For example, the inspection pattern is an all black screen (screen in which all cells on the screen are turned off)
One type, exposure time is 1 / 2s, 1 / 20s, 1 / s
When the image is switched to 200 s and imaged, three inspection pattern images are accumulated. Step S2
In 1, the luminance measurement region can be provided in the light emitting region of the cell for all cells on the image based on the image measurement result of the measurement pattern. Therefore, regarding the image of the inspection pattern, the luminance measurement region of the image is measured within the luminance measurement region of each cell. And measure the average or median. In step S22, a new image (intra-cell luminance image) in which the average value or the median value is arranged is created. The reason why the median value may be used instead of the average value is that the median value is less affected by noise than the average value. For example, even if one noise having a high brightness suddenly enters, the median value is almost unchanged, but the average value is greatly affected.

A plurality of brightness measurement areas of each cell may be set for the light emitting area of one cell. In addition, each of the brightness measurement areas is set so as to avoid the black matrix portion that is not the light emitting portion of the liquid crystal. Step S2
Since steps 1 and S22 are respectively performed on three images of the same inspection pattern with different exposure times (step S23), three in-cell luminance images are created in the first embodiment (step S24). Then, it is judged whether or not there is an image of the next inspection pattern (step S25).

In FIGS. 8A to 8C, regarding the light emitting regions (arrows W) of the R, G, and B cells, there are three luminance measuring regions for each cell, that is, WR1 to WR3 and WG1 to WG.
3 and WB1 to WB3 are set, and three in-cell luminance images E1 to E3 and a combined luminance image E are created based on three exposure times. Brightness measurement area WR1 to WR
3, the length of one side on the image of WG1 to WG3 and WB1 to WB3 is preferably composed of 3 pixels or more of the light receiving element of the image pickup means 1. Below this, the number of luminance measurement populations in one cell decreases, the reliability of the measured values decreases, and the size of one cell to be inspected and the size of one pixel of the light receiving element of the image pickup means 1 are reduced. Since they are close to each other, moire fringes (interference) are likely to occur and measurement becomes difficult.

The three in-cell luminance images E1 to E3 can be considered as images having different sensitivities. The in-cell luminance image composed of images with long exposure time is a highly sensitive image, and the luminance of cells with dark luminance is accurately measured, but the luminance value is saturated for cells with high luminance. is doing. On the other hand, an in-cell brightness image composed of images with short exposure time accurately measures the brightness of cells with high brightness, but the brightness value of cells with low brightness has a black level (dark current component). Or it is also called noise component) and the level is not changed, and the measurement cannot be performed correctly. This is because the dynamic range of the brightness of the liquid crystal display panel far exceeds the dynamic range of the image pickup means 1.

Therefore, in step S24, these 3
The images E1 to E3 are combined in the following procedure, for example, to create a combined luminance image E. Here, the image pickup means 1
The luminance gradation of has 256 gradations (0 to 255 in terms of luminance). The following procedure (step 241 and step 242) is one example of the synthesizing method.

In step 241, an image in which the luminance value of the pixel of interest is closest to 128 is selected from the three in-cell luminance images.

In step 242, if the in-cell luminance image is obtained when the exposure time is 1/2 s, the luminance of the in-cell luminance image at the 1/2 s exposure time is output. In the case of an in-cell luminance image with an exposure time of 1/20 s, (i) if the luminance is below the black level, the luminance of the in-cell luminance image with a 1/2 s exposure time is output. (Ii) If the brightness is above the black level, (1 / 20s
The brightness of the in-cell brightness image at the exposure time-black level) * the brightness ratio of 1 (10 = (1/2) / (1/20)) is output. In the case of an in-cell brightness image with an exposure time of 1/200 s, (i) if the brightness is below the black level, (brightness of the in-cell brightness image with an exposure time of 1/20 s-black level) x brightness ratio 1
(10 = (1/2) / (1/20)) is output. (Ii) If the luminance is above the black level, (1/200
s Brightness of in-cell brightness image at exposure time-black level) x brightness ratio 2 (100 = (1/2) / (1/200)) is output.

First, in step 241, an in-cell luminance image having no luminance near the black level and no luminance near saturation is selected, and in step 242, the combining process is performed. The luminance ratios 1 and 2 in the combining process correspond to the exposure time ratios.

The inspection pattern is an all-red screen in addition to the all-black screen,
If there are all-green screens or all-blue screens, composite luminance images are similarly created for the number of patterns.

By the procedure up to this point, the dynamic range of the image pickup means can be expanded and the average value or the median value in the brightness measurement area of each cell can be measured.
In S8, these values are compared with a predetermined threshold value to judge acceptability. It is assumed that the threshold is set for each cell color.

For example, when determining a bright spot defect, the average value or the median value obtained for all black images is compared with a threshold value corresponding to the minimum brightness of the bright spot defect, and the threshold value or more is determined. When the value of is obtained, the cell is determined to be defective. When inspecting whether the range of the cell brightness when displaying the brightness of the intermediate gradation is within the specified value, set two upper limit thresholds and lower limit thresholds that define good products, Just compare.

Next, the determination of bright spot defects will be described.

If there is no bright spot defect having a high brightness, the quality may be determined by comparing the average value or the median value in the brightness measurement area of each cell with a preset threshold value.
When there is a bright spot defect having high brightness, light circulates into the cells around the defective cell, and the brightness of the peripheral cells becomes high. Therefore, if a fixed threshold value is used for comparison, there arises a problem that cells around a high-luminance bright spot defect are erroneously recognized as a defect.

Therefore, in step S4, the threshold value applied to the cells around the high-luminance bright spot defect is changed to be high, for example, as follows. First, with respect to the combined luminance image E, as shown in FIG. 9, two pixels around the pixel pt to be inspected, in the present embodiment, 24 pixels around the pixel pt (from the pixel pt to all pixels of two radiuses ), It is checked whether or not the pixel is a high-luminance pixel having a luminance higher than that of the pixel pt and having a brightness equal to or higher than the luminance γ designated in advance. If the pixel is not a high luminance pixel, the luminance of the pixel to be inspected is compared with the threshold value as usual, and the quality of the cell corresponding to the pixel to be inspected is determined. The threshold value is set for each color when there are cells of different colors in the liquid crystal display panel, which is the inspection object.

On the other hand, in the case of a high brightness pixel, the threshold value increment is calculated using the threshold value increment coefficient map M shown in FIG. The threshold increment coefficient map M includes high-luminance pixels p
Centered on the map M from the position of the pixel to be inspected
Taking the value above, calculate the threshold increment by the following formula:

Threshold increment value = luminance value of high luminance pixel × value on map

In the case of the examples of FIGS. 9 to 10, the luminance of the high luminance pixel p1 on the upper side in FIG.
If the brightness of is GD, the threshold increment values are as follows. Note that in FIG. 9, for the sake of explanation, pt, p1, and p
The diagonal lines have been added to make 2 easier to understand.

[0040]     Threshold increment value (p1) = GU x 0.8β (1)     Threshold increment value (p2) = GD x 0.5α (2)

As in the case of this example, when there are a plurality of high-intensity pixels for the pixel to be inspected, the larger one of the values calculated by the equations (1) and (2) is selected. The threshold increment value. Here, the values of α and β are values to be changed depending on the type of the display device 8 (liquid crystal A or liquid crystal B, etc.) that is the inspection object, and α> β. By setting the values of α and β and the respective values on the coefficient map M to predetermined values based on experiments, it is possible to set the influence of the brightness of the high brightness pixels on the surroundings. For example, α = 0.15 and β = 0.01 are set.

After the threshold increment value is determined in this way, the brightness of the pixel to be inspected is compared with (preset threshold value + threshold increment value) to determine the pixel to be inspected. The quality of the corresponding cell is determined.

The bright spot defect determination processing in steps S3 to S8 can be summarized as the following procedure (steps 31 to 3).
It becomes 5).

In step 31, the color of the pixel to be inspected is discriminated from the coordinates of the luminance composite image.

In step 32, it is checked whether or not there is a high-luminance pixel having a luminance higher than that of the pixel to be inspected in the peripheral pixels. When there is no high-luminance pixel in the peripheral pixels, the following steps (i) to (iii) are performed depending on the color difference of the pixel to be inspected, and the determination of the pixel to be inspected is completed. (I) In the case where the color of the pixel to be inspected is R, if the luminance is larger than RTh, it is considered as a defect. Others are good. (Ii) When the color of the pixel to be inspected is G, if the luminance is larger than GTh, it is determined as a defect. Others are good. (Iii) When the color of the pixel to be inspected is B, the brightness is BTh
If it is larger, it is regarded as a defect. Others are good. If the peripheral pixels have high-luminance pixels, the process proceeds to step S33.

In step 33, it is checked whether or not the maximum brightness value of the peripheral high brightness pixels is γ or less. When the maximum brightness value of the surrounding high brightness pixels is γ or less,
The same determination processing as in the case where there is no high-luminance pixel in step 32 is performed. If the maximum brightness value of the surrounding high brightness pixels is larger than γ, the process proceeds to the next step 34.

In step 34, the threshold value increment value is calculated from each of the peripheral high-intensity pixels using the coefficient map, and the maximum value thereof is set to ThUp.

In step 35, the luminance of the pixel to be inspected is compared with (threshold value + threshold increment value) due to the difference in color of the pixel to be inspected. When the pixel to be inspected is R, the brightness is (RTh + ThU
If it is larger than p), it is regarded as a defect. Others are good. When the pixel to be inspected is G, the brightness is (GTh + ThU
If it is larger than p), it is a defect. Others are good. When the pixel to be inspected is B, the brightness is (BTh + ThU
If it is larger than p), it is regarded as a defect. Others are good.

Where RTh, GTh and BTh are
It is a preset threshold value for each color of each cell.

When there are a plurality of inspection patterns and combined luminance images, the same inspection is repeated for each combined luminance image.

Through steps S3 to S8, the presence / absence of a defect and the coordinates on the image of the defective cell are known. Then, in step S9, the position of the defective cell on the image is converted into display coordinates. By converting the position of the defective cell into display coordinates, it is convenient to output the position of the defective cell in a form independent of the inspection device. Of course, if it is not necessary to output the position of the defective cell in display coordinates, this step is unnecessary. Then, in step S10, the presence / absence of a defective cell, the luminance information of the defective cell, and the defect position are output.

According to the first embodiment, with the configuration as described above, stable inspection can be performed regardless of the presence or absence of bright cells in the periphery.

Second Embodiment In step S2 of the first embodiment, the brightness in the brightness measurement region is first measured to obtain the average value or the median value, and then the in-cell brightness images having different exposure times are combined. But,
Step S2 may be configured as follows.

First, the picked-up images with different exposure times are combined in the same procedure as steps 241 and 242 of the first embodiment to create a combined luminance image of the picked-up images. Then, in the same manner as in step S21, luminance measurement is performed on the combined luminance image in the luminance measurement area of each cell to obtain an average value or a median value, and then, similarly to step S22, the average value or the median value. To create a new image. With this method, an image corresponding to the combined luminance image obtained in step S24 can be obtained. Even with this,
The same effect as that of the first embodiment can be obtained.

Third Embodiment The conversion formula used in the conversion process of the image coordinates and the display coordinates in step S9 of the first embodiment can be obtained as follows before the inspection is started. Here, the points that are not particularly mentioned are the same as those in the first embodiment.

As shown in FIG. 11, after the image pickup means 1 is moved to the upper left corner FL of the display on which the measurement pattern is displayed, the image is picked up and the cross hatch is measured in the same manner as in step S1. When the measurement is completed, the display coordinates and the image coordinates can be associated (converted). Because the location where the image is taken is known to be the upper left corner FL of the display, as shown in FIG. 12, the imaged display coordinates at the upper left of the cross hatch are G cells of (0, 0) coordinates. This is because it is known that there is. Since the position of the camera of the image pickup means 1 can be represented by the center of the picked-up image, which is the viewpoint of the camera, what represents the center position of the image at this time in display coordinates is ( ^D X ₁ , ^D Y ₁ ). To do. Further, the absolute position of the image pickup means 1 obtained from the measuring means 4 at this time is ( ^W X ₁ , ^W Y ₁ ).
And

It is assumed that the conversion equation between the absolute coordinates and the display coordinates can be performed by the following affine transformation. ( ^D X, ^D Y) ^T = A ( ^W X, ^W Y) ^T + b (3)

Here, A is a 2 × 2 matrix showing rotation, magnification and distortion in the XY directions, and b is a two-dimensional vector showing a translation component. T indicates the transposition of the vector. In this case, in the present embodiment, if there are at least three combinations of the display coordinates and the absolute coordinates, the coefficients A and b of the affine transformation can be calculated, but in the present embodiment,
As shown in FIG. 11, by capturing the four corners of the display image (upper left corner FL, upper right corner FR, lower right corner RR, lower left corner RL), A and b are combined from the combination of the four points.
To calculate. When calculating from the combination of coordinates of three or more points, regression calculation such as the least squares method is performed on the form of formula (3).

By calculating this affine transformation formula in advance, in step S1 of the first embodiment,
The conversion formula between the image coordinates and the display coordinates can be calculated accurately. Further, in step S1, the center positions P of the plurality of cells C are calculated as shown in FIG. 5, but at the same time, the absolute position ( ^W X _M , ^W Y _M ) ^T of the image pickup unit 1 is measured by the measuring unit 4. get. This absolute position can be calculated using equation (3) using the coordinates ( ^D X _M , ^D Y _M ) ^T on the display screen.
Convert to. If the position of the image pickup means 1 and the conversion formula are extremely accurate, these display coordinates will be the display coordinates of the point seen at the center position of the current image. Since the crosshatch is displayed at a preset location, the measurement point position on the crosshatch that should be the closest to the coordinates ( ^D X _M , ^D Y _M ) ^T indicating the center position of this screen, that is, the measurement point display coordinates _{^{(D X P, D Y P}} ) can be calculated ^T. Also,
In the image measurement step S1, since it calculates the magnification of the display coordinates and image coordinates, to calculate by multiplying the magnification difference _{^{(D X M, D Y M}} ) T and _{^{(D X P, D Y P}} ) T so,
Image coordinates can also be calculated. The position of this image coordinate will be called the predicted position of the measurement point.

However, actually, the measurement point on the crosshatch is not always measured at the position of the predicted image coordinate. This is because there are mechanical elastic deformations of the moving means 3 and measurement errors of the measuring means 4, so that the measurement point of the crosshatch may be actually measured at a shifted position on the image.
Therefore, the image coordinates of the measurement point closest to the predicted position on the image are ( ^D _XP ,
It is assumed that it corresponds to ^D Y _P ) ^T. In order for this assumption to hold, the amount of deviation due to an error or the like needs to be 1/2 or less of the crosshatch interval. On the contrary, when the upper limit of the shift amount can be estimated, the assumption is always valid if the crosshatch interval is set to be equal to or more than the upper limit value. If the image coordinates and the display coordinates can be matched at this one point, all the display coordinates of the measurement points whose display positions are predetermined are determined. Therefore, the correspondence between the display coordinates and the image coordinates is accurate at all the measurement points. It is possible to remove the influence of measurement error. Even in the conversion formula of image coordinates and display coordinates,
If the same affine transformation equation as the equation (3) is used, the transformation equation of the image coordinates and the display coordinates can be calculated with the combination of the coordinates of three or more points.

If the conversion formula obtained in this way is used in step S9 in the first embodiment, output in display coordinates becomes possible.

Fourth Embodiment In the third embodiment, the conversion formula of the image coordinates and the display coordinates is configured by using the affine conversion formula. However, if the following is performed, a larger number of points, that is, almost all measurement values are measured. By using the points, it is possible to construct a conversion formula with higher accuracy than the affine transformation.

In this example, since the measurement points are arranged in a grid pattern, a small region SR is created for each set of 9 points as shown in FIG. Then, the relationship between the image coordinates and the display coordinates is regressed by a quadratic curve for each area, and the parameter is stored for each area. The shape of the quadratic curve is, for example, as in the following expressions (4) and (5).

[0064]

[Equation 1]

Here, (X, Y) are coordinates before conversion, (U, V) are coordinates after conversion, and a _ij and b _ij are 2
This is the parameter of the next curve. Also, a and b are 6 each
There is one. When obtaining the conversion formula of the image coordinates from the display coordinates, (X, Y) is set as the display coordinates, and (U,
V) is the image coordinates, and the parameters are calculated from the combination of the coordinates of 9 points. Further, when a conversion formula from the image coordinates to the display coordinates is necessary, the parameters are calculated with (X, Y) as the image coordinates and (U, V) as the display coordinates. These parameters are calculated for each small area, and both the image coordinates and the display coordinates of the point that is the center of the small area are saved. In the actual conversion, when the image coordinates are given, it is first calculated which small area belongs to which measurement point is close, and based on the equations (4) and (5) using the parameters of the small area to which it belongs. And calculate the conversion.

In the affine transformation described above, the transformation calculation is constant regardless of the position on the screen. However, there is a characteristic that the distortion of the lens forming the imaging means is small in the center of the screen and large in the peripheral part. That is, the amount of distortion varies depending on the position on the screen. In such a case, conversion calculation cannot be performed accurately with a constant calculation method on the screen. In this way, by dividing multiple measurement points into small areas and performing conversion calculations by changing the regression calculation parameters for each small area, the amount of distortion that differs depending on the position on the screen can be corrected, resulting in a more accurate display. It is possible to obtain a transformation formula of coordinates and image coordinates.

In the case of the third embodiment, when the image coordinates and the display coordinates of the plurality of measurement points are obtained in step S1, the conversion formulas for the display coordinates and the image coordinates are calculated. A more accurate cell position can be calculated by calculating the luminance measurement area based on the display coordinates of all cells to be inspected and applying the conversion formula to obtain the image coordinates of the luminance measurement area.

Fifth Embodiment In the first embodiment, RTh used in step S2,
You may perform the calculation method of the threshold value set up beforehand, such as GTh and BTh, as follows.

First, when a combined image of in-cell luminance images is obtained, once for all cells in the screen, R _BASE , G _BASE , B which are average values or median values classified by color.
Calculate _BASE . Normally, the number of defective cells should be much smaller than the number of cells in the entire screen, and thus these values can be considered as typical brightness values of normal cells. In addition, normally, the normal cell luminance value differs depending on the color due to the difference in color sensitivity of the image pickup means and the transmission characteristic of the color filter forming the liquid crystal display device as the inspection object. Therefore, the threshold value is configured by adding a reference value for detecting an abnormal value to the value considered to be the luminance value of these normal cells. That is, it is possible to configure the _{RTh = R BASE + R STD ·} VALUE GTh = G BASE + G STD · VALUE BTh = B BASE + B by an _{STD · VALUE,} more accurately inspect a possible threshold. Where R _{STD · VALUE} and G _{STD · VALUE}
And B _{STD · VALUE} are thresholds for detecting abnormal values corresponding to R, G, and B, respectively.

If the sensitivity characteristic according to the color of the image pickup means is known in advance, the luminance may be corrected by multiplying the result of the luminance measurement obtained in step S2 by a coefficient based on the sensitivity characteristic. By doing so, an accurate inspection that is not affected by the color of the cell can be realized.

[0071]

According to the first aspect of the present invention, when comparing the brightness average value or the median value of each cell with the threshold value, when a cell with high brightness exists near the cell to be inspected, Since the threshold value is increased according to the brightness value of the cells in the vicinity thereof, even if there is a bright spot defect cell having high brightness, the inspection can be performed accurately and accurately.

According to the second aspect of the present invention, with respect to the luminance measurement, the luminance measurement is performed a plurality of times in each of the luminance measurement areas while changing the exposure time of the image pickup means, and the obtained luminance measurement results are combined. Then, since the brightness in each brightness measurement region is calculated, it is possible to solve the shortage of the dynamic range of the imaging means, and it is possible to inspect both dark defects and bright defects at the same time.

According to the third aspect of the present invention, the specific coordinate position on the display screen, the position on the image obtained by the image pickup means relating to the coordinate position, and the absolute position data of the image pickup means at that time are stored. Using multiple sets, calculate the absolute position of the imaging means and the conversion formula on the display screen in advance,
At the time of inspection, a deviation between the absolute position of the image pickup means and the expected position on the image of the measurement pattern obtained by using the conversion formula and the position on the image of the actual measurement pattern is measured, and this deviation amount is used. A new conversion formula for image coordinates and display coordinates is calculated using this conversion formula, and the position of the inspected cell on the image is converted to the display coordinate position and output using this conversion formula. The accurate display coordinate position can be output without being affected by the measurement error of the measuring means based on.

According to the invention described in claim 4, in the position measurement pattern, a plurality of cells at predetermined positions on the display screen are turned on, and the positions of the cells at a plurality of points are measured.
In order to calculate the position of each cell by dividing them into small areas and performing a curve regression of the relationship between the image coordinates and the display coordinates for each small area to correct the distortion amount of the image of the image pickup means. The brightness measurement area of the cell can be set accurately.

According to the fifth aspect of the present invention, in the creation of the threshold value, an average value or a median value is calculated for each cell having a different color for the entire inspection region, and the average value or the median value is used as a reference value. Since a threshold value is created for each cell having a different color by adding, it is possible to accurately inspect even if there is a difference in sensitivity due to the color of the image pickup means or a difference in light emission amount due to the color of the cell to be inspected. it can.

According to the sixth aspect of the present invention, in the luminance measurement, the luminance measurement result is corrected for each cell having a different color by multiplying it by a coefficient based on the sensitivity characteristic of the image pickup means. It is possible to correct the difference in sensitivity due to color and perform an accurate inspection.

According to the invention described in claim 7, image pickup means for picking up an image of the display screen, display screen control means for controlling the screen of the display screen, moving means for holding the image pickup means and enabling movement. A measuring unit that measures the absolute position of the image capturing unit and an image of the display screen that displays the position measurement pattern captured by the image capturing unit are processed to measure the position of each cell that constitutes each pixel, By processing the image of the display screen on which the inspection pattern is displayed, which is imaged using the image pickup means, the brightness measurement areas are provided at the positions of the respective measured cells, and the brightness measurement is performed in each of the brightness measurement areas. Calculate the average value or median value of the obtained brightness measurement results, compare with the threshold value, inspect each cell, and determine the brightness average value or median value of each cell. When there is a cell with high brightness in the vicinity of the cell to be inspected when comparing the two values, an image processing unit configured to increase the threshold value according to the brightness value of the cell in the vicinity, Since the control means for controlling the entire flow of the inspection is provided, it is possible to perform the inspection accurately and accurately even if there is a bright spot defect cell having high brightness.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a display screen inspection device according to a first embodiment of the present invention.

FIG. 2 is a flowchart for explaining a display screen inspection method according to the first embodiment.

FIG. 3 is a diagram for explaining a measurement pattern according to the first embodiment.

FIG. 4 is an enlarged view of a main part of FIG.

FIG. 5 is a diagram for explaining center positions of a plurality of cells measured from the measurement pattern according to the first embodiment.

FIG. 6 is a diagram for explaining a method of measuring the center positions of a plurality of cells from the measurement pattern according to the first embodiment,
FIG. 6A is a diagram for explaining the projection process, FIG. 6B is a diagram for explaining window setting, centroid measurement and line extraction, and FIG. 6C is a diagram for explaining intersection detection.

FIG. 7 is a flowchart for explaining step 2 of the display screen inspection method according to the first embodiment.

FIG. 8 is a diagram for explaining a procedure for creating an in-cell luminance image and a combined luminance image in step 2 of the display screen inspection method according to the first embodiment.

FIG. 9 is a diagram for explaining the presence / absence of a high-intensity cell around a pixel to be inspected in the display screen inspection method according to the first embodiment.

FIG. 10 is a diagram for explaining a threshold value increment coefficient map for calculating a threshold value increase value due to a high brightness cell in the display screen inspection method according to the first embodiment.

FIG. 11 is a diagram for explaining a method of obtaining a conversion formula between a display coordinate and an absolute coordinate in the display screen inspection method according to the third embodiment.

FIG. 12 is a diagram showing an upper left corner of a display image.

FIG. 13 is a diagram for explaining a method of obtaining a conversion formula between image coordinates and display coordinates in the display screen inspection method according to the fourth embodiment.

FIG. 14 is a configuration block diagram for explaining a conventional liquid crystal display panel inspection method.

FIG. 15 is a diagram for explaining a relative positional relationship between display pixels of a liquid crystal display panel and photocells forming an area type photosensor built into a CCD camera in a conventional liquid crystal display panel inspection method.

FIG. 16 is a diagram for explaining alignment of the liquid crystal display panel in the conventional liquid crystal display panel inspection method.

[Explanation of symbols]

1 image pickup means, 2 display screen control means, 3 moving means, 4 measuring means, 5 image processing means, 6 control means, 7 determination results, 8 display devices.

Claims (7)

[Claims] 1. The position of each cell constituting each pixel is measured by using an image of a display screen on which a position measurement pattern is displayed, which is imaged by an imaging means, with respect to a predetermined display area of the display screen. Using the image of the display screen on which the inspection pattern is displayed, which is picked up by the image pickup means, a luminance measurement region is provided at each measured cell position, and the luminance measurement is performed in each luminance measurement region. A method of inspecting each cell by calculating the average value or median value of the obtained brightness measurement results and comparing with the threshold value, and comparing the brightness average value or median value of each cell with the threshold value. When there is a cell with high brightness in the vicinity of the cell to be inspected, the display screen inspection method for increasing the threshold value according to the brightness value of the cell in the vicinity. 2. Regarding the luminance measurement, the luminance measurement is performed a plurality of times in each luminance measurement region while changing the exposure time of the image pickup means, and the obtained results of the luminance measurement are combined to obtain each luminance measurement region. The method for inspecting a display screen according to claim 1, wherein the brightness in the inside is calculated. 3. An image pickup means using a plurality of sets of data of a specific coordinate position on a display screen in advance, a position on the image obtained by the image pickup means concerning the coordinate position, and an absolute position of the image pickup means at that time. Absolute position and the conversion formula on the display screen are calculated in advance, and at the time of inspection, the absolute position of the imaging means and the expected position on the image of the measurement pattern obtained using the conversion formula, and the actual measurement pattern. The displacement of the position on the image of the cell is measured, the conversion formula of the image coordinate and the display coordinate is newly calculated using this displacement amount, and the position of the inspected cell on the image of the display coordinate position is calculated using this conversion formula. The display screen inspection method according to claim 1, wherein the display screen is converted and output. 4. Regarding the position measurement pattern, a plurality of cells at predetermined positions on a display screen are lit, the positions of cells at a plurality of points are measured, and these are divided into small areas,
2. The display screen inspection method according to claim 1, wherein the position of each cell is calculated by correcting the distortion amount of the image of the image pickup means by performing a curve regression on the relationship between the image coordinates and the display coordinates for each small area. 5. In the creation of the threshold value, an average value or a median value is calculated for each cell having a different color for the entire inspection area, and a reference value is added to the average value or the median value to make the color different. The display screen inspection method according to claim 1, wherein a threshold value is created for each cell. 6. The display screen inspection method according to claim 1, wherein, for the luminance measurement, the luminance measurement result is corrected for each cell having a different color by multiplying it by a coefficient based on the sensitivity characteristic of the image pickup means. 7. An image pickup means for picking up an image of a display screen, a display screen control means for controlling the screen of the display screen, a moving means for holding the image pickup means to allow movement, and an absolute position of the image pickup means. An image of the display screen on which the position measurement pattern is displayed, which is imaged using the measuring unit and the image capturing unit, is processed to measure the position of each cell that constitutes each pixel, and the image is captured using the image capturing unit. By processing the image of the display screen on which the inspection pattern is displayed, a brightness measurement region is provided at each measured cell position, and the brightness measurement is performed in each brightness measurement region. Calculate the average value or median value and compare it with the threshold value to inspect each cell, and compare the brightness average value or median value of each cell with the threshold value. When there is a high-luminance cell in the vicinity of the target cell, the image processing means configured to raise the threshold value according to the luminance value of the neighboring cell and the control of the entire inspection flow are performed. A display screen inspection device comprising control means.