JPH11132750A - Point defect detecting device and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make securable the flaw detection sensitivity without increasing the number of cameras and the number of picture elements by specifying the approximate position of the point defect of a component with a camera, and detecting the point defect from the density data of the picture elements of gray level images of the component. SOLUTION: The light of a light source 3 is radiated from the back of a liquid crystal panel 1, and it is photographed by a camera 2. The camera 2 has an image pickup element and feeds variable-density signals for individual picture elements to a computer 4. An optical path variable rotary plate 6 driven by a drive motor 77 is provided in the front of the camera 2, and it is rotated 360 deg. during the exposure time to drift the relative positional relation between the image pickup element and the liquid crystal panel 1 for multiplex exposure. The position of the picture element with the maximum density within the lump of picture elements with the density brighter than the density threshold value is made the flaw approximate position, an intermediate-magnification lens 5 is moved by a moving mechanism 76 based on it, and the detailed position is obtained by the camera 2. The intermediate- magnification lens 5 is moved to all picture element flaw positions, images are obtained after the lens 5 is focused by a focusing mechanism 75, and picture element defects are judged by the computer 4.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the inspection of point defects on the surface of electronic components such as liquid crystal panels, CRT panels, PDP panels (plasma display panels) used in the field of electronic equipment and the like in a production line. The present invention relates to a point defect detection apparatus and method for detecting a defect based on density data of each pixel of a grayscale image corresponding to a point defect on the surface obtained by imaging an object to be inspected with an imaging element.

[0002]

2. Description of the Related Art Conventionally, screen inspection has been carried out either by manual inspection or by image processing by an automatic machine. In the case of manual operation, even if the type of liquid crystal is changed, it can easily cope with the change, and the rise is quick, but it takes time to specify the detailed position of the defect, and there is a drawback that the throughput is poor. On the other hand, in the case of an automatic machine, if the model is changed, much time is required for adjustment, but there is an advantage in that the identification information of the defect position is quickly recognized. However, with the recent increase in the definition of liquid crystal panels, the number of pixels of the image sensor has been relatively reduced with respect to the number of pixels of the liquid crystal panel.

[0003] When imaging a liquid crystal panel under the condition that the number of pixels of the liquid crystal panel is 500 pixels in the horizontal direction and 500 pixels in the vertical direction, and the number of pixels of the liquid crystal panel is 1000 pixels in the horizontal direction and 1000 pixels in the vertical direction, one pixel of the liquid crystal panel is Pixels are assigned two horizontal pixels and two vertical pixels. In this case, defect detection sensitivity and defect position identification accuracy are ensured. However, under the same conditions, the number of pixels of the liquid crystal panel is 2000 pixels
Under the condition of 2,000 pixels in the vertical direction, only one half pixel in the horizontal direction and one half pixel in the vertical direction are allocated to one pixel of the liquid crystal panel. In this case, the defect detection sensitivity and the defect position identification accuracy are not secured.

As shown in FIGS. 10 (a) and 10 (b), a pixel defect 5 of a liquid crystal panel is
The case where 1 is one horizontal pixel and one vertical pixel will be taken as an example. Differences occur in the detected density data depending on the relative positions of the image sensor and the liquid crystal panel. The density data received by the image sensor is 25
When it is possible to express in six gradations, it is assumed that a dark part is close to 0 and a bright part is close to 256. Pixel defects of liquid crystal panel are
In the case of black lighting, the defect shines brightly. The following detected concentration data is obtained. [1] In the state of FIG. 10A, the density data of the pixels assigned to the background portion other than the defect is around 30, and the density data of the pixels assigned to the defective portion is 150. [2] In the state of FIG. 10B, the density data of the pixels assigned to the background portion other than the defect is around 30, and the density data of the pixels assigned to the defective portion is 60.

[0005] Thus, various relative positions occur.
The smaller the number of assigned pixels, the greater the fluctuation of the obtained density data. In the above case, considering the data of the background part separately, in the state of FIG.
20 and in the state of FIG. 10B, the defect density data is 30, which is a four-fold difference. This is due to the existence over four pixels. For example, assuming that the density threshold for determining a defective portion is 90, detection is possible in the state of FIG. 10A, but not in the state of FIG. 10B. Even with the same defect of the liquid crystal panel, a difference occurs in the detection sensitivity due to a difference in the relative positional relationship with the image sensor. This has been dealt with by increasing the number of cameras or the number of pixels of the image sensor.

[0006]

However, there are limitations on the physical size of the camera and limitations on the number of pixels of the image sensor, and it is becoming difficult to secure defect detection sensitivity and defect position identification accuracy. SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a point defect detection apparatus and method capable of securing defect detection sensitivity and defect position identification accuracy without depending on the number of cameras and the number of pixels of an image sensor. .

[0007]

In order to solve the above-mentioned problems, the present invention is configured as follows. First of the present invention
According to the aspect, an image of a component is captured by an image sensor, and a defect approximate position specifying camera that roughly specifies a defect position of a point defect in the captured image, and a camera between the image sensor and the component of the camera. It is arranged rotatably in the optical axis of, and, during imaging, by rotating around the rotation axis, the relative optical axis position of the component and the imaging device is continuously shifted, and the relative optical axis position An optical path changing rotary plate that can capture the image of the component with the misalignment into the image sensor of the camera, and a lens for specifying a defect detailed position by capturing the image obtained by partially enlarging the component into the camera. Identifying the approximate position of the point defect of the component with the camera and detecting the point defect from the density data of each pixel of the gray image of the component captured by the camera through the lens according to this information Identifying the defect detailed position Te provides a defect detection device that characterized.

According to a second aspect of the present invention, there is provided the point defect detecting apparatus according to the first aspect, wherein the component is a liquid crystal panel, and a point defect in a liquid crystal screen of the liquid crystal panel is inspected. According to the third aspect of the present invention, in the image obtained by the camera through the lens, from the color, shape, and height, whether the image is a pixel defect on the screen itself or a foreign matter contamination defect on the screen itself A second aspect of the present invention is to provide a point defect detection apparatus for distinguishing whether foreign matter is attached to a screen surface.
According to a fourth aspect of the present invention, there is further provided a moving mechanism for moving the lens in a direction orthogonal to the optical axis of the component, and moving the lens according to the information on the approximate position of the defect to thereby image the defect. The point defect detection device according to any one of the first to third aspects, wherein the point defect detection device is configured to specify the defect detailed position.

According to the fifth aspect of the present invention, during imaging, the optical path changing rotary plate, which is arranged in the optical axis between the image sensor and the component of the camera for specifying the approximate position of a defect, rotates the optical path changing rotary plate around the rotary axis. Then, the relative optical axis position between the image sensor and the component is continuously shifted, and the image of the component whose optical axis position is relatively shifted is taken into the image sensor of the camera, and the image pickup is performed. Based on the image captured by the element, a defect approximate position of a point defect in the image is specified, and based on the defect approximate position of the specified point defect, an image obtained by partially enlarging the part is enlarged. A part of the image of the component is enlarged and captured by the camera through a lens that can be captured by the camera, and the point defect is detected from density data of each pixel of a grayscale image of the component captured by the camera through the lens. And defect details Identifying a position to provide a failure detection method that is characterized in.

According to a sixth aspect of the present invention, there is provided the point defect detecting method according to the fifth aspect, wherein the component is a liquid crystal panel, and a point defect in a liquid crystal screen of the liquid crystal panel is inspected. According to the seventh aspect of the present invention, in an image obtained by the camera through the lens, from the color, shape, and height, a pixel defect on the screen itself, a foreign matter mixed defect on the screen itself, or a foreign matter A method for detecting a point defect according to a sixth aspect of the present invention, which distinguishes whether the surface defect is attached to the screen surface. According to an eighth aspect of the present invention, there is further provided a moving mechanism for moving the lens in a direction orthogonal to an optical axis with the component,
A point defect detection method according to any one of the fifth to seventh aspects, wherein the lens is moved in accordance with the information on the approximate defect position to capture an image of the defect and specify the detailed defect position.

[0011]

Embodiments and Examples of the present invention will be described below in detail with reference to the drawings. FIG.
1 illustrates a configuration of a liquid crystal screen defect detection device as an example of a point defect detection device according to a first embodiment of the present invention. A light source 3 from the back of the liquid crystal panel 1 as an example of a component
And the camera 2 shoots an image. The camera 2 has an image sensor 50, and a gray-scale signal for each pixel is obtained, which is sent to the computer 4 in a digitized form. This computer 4 is, as shown in FIG.
Inside the memory, a memory 4a, an operation unit 4b connected to the memory 4a, and a program memory 4 connected to the operation unit 4b
c. In the computer 4, for example, 8-bit (0 to 255) density data is stored in a form corresponding to the rows and columns of the pixels. In the computer 4, a program for executing the following processing is set in the program memory. An optical path changing rotary plate 6 made of glass having a thickness of, for example, about 5 mm is provided in front of the camera 2 so as to be rotatable by driving a drive motor 77, and the image is formed by rotating once during the exposure time of the image sensor. Element 5
Multiple exposure is performed while shifting the relative positional relationship between 0 and the liquid crystal panel 1. In order to specify the detailed position of the defect detected by the information from the camera 2, an enlarged intermediate lens 5 composed of a single lens such as a magnifying glass is used.
The magnifying intermediate lens 5 automatically adjusts the focus position of the horizontal moving mechanism 76 and the magnifying intermediate lens 5 that enable the magnifying intermediate lens 5 to move along the liquid crystal surface of the liquid crystal panel 1, and moves from the focused position. A focus positioning mechanism 75 capable of detecting the height of a foreign substance or the like is provided. Rotating plate 6
The drive of the drive motor 77, the horizontal moving mechanism 76, and the focus position adjusting mechanism 75 is performed under the control of the computer 4. Therefore, it is possible to take an image with the camera 2 through the magnifying intermediate lens 5 and obtain a partially magnified image. If the camera 2 is not a color camera, the bandpass filter 9
Is also prepared. First, a detection procedure of the liquid crystal screen defect detection device according to the first embodiment of the present invention will be described with reference to FIG.

In step S1, the liquid crystal panel 1, which is an example of the device under test, is turned on. Here, the description is limited to black lighting. Of course, the application to other lighting is easy.
In step S2, an image on the liquid crystal panel 1 is captured by the camera 2. FIG. 3 shows a liquid crystal panel on the image sensor 50 when the angular position of the optical path changing rotary plate 6 is the position 7 shown in FIG. 3A and the position 8 shown in FIG. Light-receiving shifts (b) and (d) of one pixel defect 51
Is shown. FIG. 3A shows the light receiving shift caused by the condition shown in FIG.
FIG. 3B shows the light reception shift under the conditions shown in FIG. 3C. As shown in FIGS. 3A and 3B, when the rotating plate 6 is not provided, the image sensor 50 that receives light at the position of the pixel 51 is arranged with the rotating plate 6 and the rotating plate 6 is disposed. If it is located at the angular position indicated by 7), it is located at the position 53 of the image sensor 50, and the light receiving position is shifted as compared with the case where the rotating plate 6 is not provided. FIG.
As shown in FIG. 3C and FIG. 3D, when the rotating plate 6 is not provided, the image pickup device 50 that receives light at the position of the pixel 51 is provided with the rotating plate 6 and the rotating plate 6 shown in FIG. 8 is located at the position 54 of the image sensor 50, and the light receiving position is shifted as compared with the case where the rotary plate 6 is not provided. Here, the image of the liquid crystal panel 1 is set to be received by the image sensor 50 with the optical path shifted by half a pixel of the image sensor 50.

Here, the angle of the rotating plate 6 with respect to the optical path deviation will be described with reference to FIG. When the material of the rotating plate 6 is glass, assuming that the glass thickness is d, the optical path deviation is w, the incident angle is a ₁ , the internal refraction angle is a ₂ , and the coefficients are n ₁ and n ₂ ,
It looks like this:

N ₁ sina ₁ = n ₂ sina ₂ w = d tana ₂ n ₁ = 1, n ₂ = 1.45 For example, when the incident angle: a ₁ = 4 °, the glass thickness: d
= 2 mm, the optical path shift: w = 96 μm. Here, the pixel of the liquid crystal panel 1 is 200 μm / pixel,
It is assumed that one horizontal pixel and one vertical pixel of the image sensor 50 are assigned to this one pixel. When the optical path from the liquid crystal 1 is shifted by about 100 μm by the optical path changing rotating plate 6, the light is received by the image pickup device 50 shifted by 画素 pixel. When the exposure time of the image sensor 50 is 200 ms and the rotation speed of the optical path changing rotary plate 6 is 200 ms, the optical path is shifted by about 100 μm in all directions. Adjustment of the amount of received light
And the exposure time.

FIGS. 5A and 5B show the light receiving position when the optical path is shifted by the rotation of the rotary plate 6. FIG. The original light receiving position 51 becomes the light receiving position 52 by being shifted by 1/2 pixel. Here, the following detected concentration data is obtained. However, the density data of the pixel assigned to the defective portion is the pixel having the largest assigned area, and has the largest density data. For example, FIG.
In the state (1), as an example, the density data of the pixels assigned to the background portion other than the defect is around 30, and the density data of the pixels assigned to the defective portion is 150. In the state of FIG. 5B, as an example, the density data of the pixel assigned to the background portion other than the defect is about 30 and the density data of the pixel assigned to the defect portion is 15%.
0. Here, assuming that the density threshold value for determining whether or not a defective portion is 90 is, for example, 90, the density data of the defective portion is higher than the threshold value and the background portion is lower than the threshold value in the state of FIG. Therefore, the defective portion can be detected from the background portion. Even in the state shown in FIG. 5B, since the density data of the defective portion is higher than the threshold value and the background portion is lower than the threshold value, the defective portion can be detected from the background portion.

By the way, even in the same defect of the liquid crystal panel, a difference in detection sensitivity occurs in the conventional example due to a difference in a relative positional relationship with the image pickup device. However, in the liquid crystal screen defect detection device according to the first embodiment, the fluctuation of the obtained density data is small. That is, when imaging the liquid crystal panel 1, the defect detection sensitivity is ensured even if the relative position fluctuates and the assignment of the pixels of the image sensor 50 to one pixel of the liquid crystal panel 1 is small. Since the optical path changing rotary plate 6 is rotated by 360 degrees, the optical path between the liquid crystal panel 1 and the image sensor 50 can be forcibly changed by 360 degrees, and the defective portion 51 spreads to 52 and is imaged. , Defect detection sensitivity is ensured.

In step S3, the position of the pixel having the largest density data among the clusters of pixels having the density data brighter than the density threshold value is determined as the position of the defect. FIG.
(A) shows the captured density data 20. The density data 20 includes a liquid crystal panel 21 and two pixel defects 22 and 23 of the liquid crystal panel 21. FIGS. 6B and 6C
Indicates density data of the pixel defects 22 and 23. here,
Only density data having a density threshold value of 90 or more is shown.
When the position of the pixel having the largest density data among the cluster of pixels having the density data brighter than the density threshold value 90 is set as the defect position, the pixel defect 22 shown in FIG. In the pixel defect 23 of (c), the pixel 25 is located at the approximate defect position.

In steps S4 and S5, the purpose is to obtain a detailed position by the camera 2 by moving the magnifying intermediate lens 5 based on the approximate position of the defect. The position of the magnifying intermediate lens 5 is accurately monitored by the horizontal moving mechanism. In FIG. 6A, the liquid crystal panel 1
Four locations x _1, x _2, to detect the position of the outer peripheral pixels y _1, y _2. The magnifying intermediate lens 5 is moved to a position where the approximate position of the defect is reflected. FIG. 6D shows an image obtained by the camera 2 through the magnifying intermediate lens 5. The enlarged intermediate lens 5 is aligned with the center of the pixel defect 22, and the area 26 is shown enlarged. Also at this time, the position of the moved intermediate lens 5 is monitored by a stage position detection mechanism or the like. FIG. 7A shows an image portion by the magnifying intermediate lens 5 at the left end of the liquid crystal panel. The background portion 30 is called a black matrix and is a portion that does not transmit light. Each of the color filter portions 31, 32 and 33 transmits only light of red (hereinafter abbreviated as "R"), green (hereinafter abbreviated as "G"), and blue (hereinafter abbreviated as "B"). Part. Even in the case of black lighting, the color of the color filter can be recognized by enlarging. Here, the position 35 of the R pixel which is the outer peripheral pixel is detected. Taking the color liquid crystal panel as an example, the magnification through the magnifying intermediate lens 5 is set so that the image sensor has a high magnification enough to allocate 100 pixels (10 horizontal pixels × 10 vertical pixels) to the RGB pixels of the color liquid crystal. . FIG. 7B shows a part of an image of the pixel defect 22 of the liquid crystal panel by the enlarged intermediate lens 5. A portion 32 that transmits only G, which should not be lit originally
36, and a portion 37 of the portion 33 that transmits only B is illuminated. By recognizing the position of the magnifying intermediate lens 5 by the moving mechanism 76, the outer peripheral position on the image at that time, and the defective pixel position, the computer 4 determines which pixel is defective from the known number of pixels of the liquid crystal panel. to decide. This determination operation is exemplified below.

The number of pixels of the liquid crystal panel is determined by the panel. For example, in the case of VGA, it is 640 × 480 pixels {× 3 (RGB)}. Referring to FIG. 6, since the known number of pixels m is included in the horizontal axes X ₁ and X ₂ , (X ₂
−X ₁ ) / m = pixel pitch p. Assuming that the coordinate on the horizontal axis of the position of the defective portion is x, (x−X ₁ ) / p is calculated to determine the number of the pixel from the left. From the area of the shining part, the defective pixel 22 is G1 / 2 · B1 / 3.
It is called a continuous bright spot. Next, the magnifying intermediate lens 5 moves to the position of the pixel defect 23, captures an image, and similarly determines which pixel is defective.

When the camera 2 is not a color camera, pixels are specified by the method described below. In the liquid crystal panel 1, the arrangement of the color filters is determined in advance depending on the type of the panel. For example, RGB, GRB, B
GR and the like. Therefore, if one of the colors can be specified, the colors of all the pixels can be specified. By setting a bandpass filter 9 that transmits any of the RGB colors more than the other colors before and after the lens 5 (for example, at a position as shown in FIG. 1) as viewed from the camera 2, the basic transmittance can be reduced. The color of the color filter can be specified by the difference. For example,
When the filter 9 is not provided, the density data of the normal pixel is set as follows. R pixel: 200, G pixel: 200, B pixel: 200 When the transmittance of G is 100% and the transmittance of RB is 50%,
The density data of a normal pixel shown below is used. R pixel: 100, G pixel: 200, B pixel: 100 By detecting the difference between the density data, the position of the RGB filter can be confirmed. Further, it is possible to correct the RGB density data based on the known transmittance balance. When the density data is corrected with the transmittance, the process returns to the following pixel density data. R pixel: 200, G pixel: 200, B pixel: 200

In this way, under the control of the computer 4, the magnifying intermediate lens 5 is moved to the positions of all the pixel defects by the moving mechanism 76, and the focal position is adjusted by the focal position adjusting mechanism 75. And the computer 4 determines which pixel is defective. According to the first embodiment, the defect detection and the defect position identification are shared, so that the number of cameras and the number of pixels of the image sensor can be increased irrespective of the high definition of the component for detecting the defect, for example, the liquid crystal panel. Therefore, the defect detection sensitivity and the defect position specifying accuracy can be ensured. In other words, by inserting the defect detail position specifying lens 5 as an intermediate lens into the optical path, the pixel resolution can be considerably increased and a margin can be provided, so that defect detection sensitivity and defect position specifying accuracy can be ensured. Becomes possible.

Next, a description will be given of a liquid crystal screen defect detecting apparatus according to a second embodiment of the present invention. In the second embodiment, in the image obtained by the camera 2 through the magnifying intermediate lens 5, based on the color, shape, and height, whether the image is a pixel defect on the screen itself or a foreign matter contamination defect on the screen itself ,
A distinction is made as to whether foreign matter is attached to the screen surface. FIG. 8 shows a procedure of a defect detection process according to the second embodiment, and FIG. 9 shows a diagram for explaining the process. Here, as shown in FIG. 9B, the processing is advanced by the enlarged intermediate lens 5 and the camera 2.

After the operations of steps S1 to S5 in FIG. 2 are performed to specify the defect detailed position, in step S10 in FIG. 8, the color of the defective portion is extracted and color identification is performed. As an example, it is assumed that as a result of the processing in steps S1 to S5, 41, 42, and 43 in FIG. Here, the density data shown is the corrected density data after the RGB pixels can be specified. FIG.
FIG. 9C shows a confirmation line 45 of the defect 43 shown in FIG.
(The “confirmation line” referred to here is a line for drawing a line in the center of the target of interest to indicate the pixel density on that line.) The density profile is shown. . In FIG. 9C, it is determined that the density data of G on the G filter (green filter) is larger than the respective density data of RB. FIG. 9D shows the density profile of the confirmation line 44 of the defect 41 shown in FIG. 9A. In FIG. 9D, RGB on the G filter
It is determined that there is no large difference between the respective density data. FIG. 9E shows a density profile of the confirmation line 46 of the defect 42. In FIG. 9E, it is determined that there is no large difference between the respective density data of RGB on the B filter.

Next, in step S14, it is determined whether the color is different from the color of the original color filter, in other words, whether it is a pixel defect or a foreign matter, and it is determined whether the foreign matter is a foreign matter.
An example of the determination will be described below. Here, the foreign substances are R, G,
The difference between B and the liquid crystal pixel is determined by utilizing the fact that the density value of the color corresponding to the original color filter is increased while the density value of B is substantially the same. 1) If the ratio between the density data of the largest color and the density data of the second largest color is 1.5 or more, it is determined as a pixel defect. However, the condition is that the largest color is the same color as the filter. 2) If the ratio between the density data of the largest color and the density data of the second and third largest colors is 1.49 or less, it is regarded as a foreign substance.

After the color of the defective portion is extracted in step S10, the shape of the defective portion is extracted in step S11. Then, in step S12, it is determined whether or not the original color filter portion protrudes from the original color filter portion. If so, it is determined in step S13 that the color filter is abnormal. If not, step S14
Proceed to. Specifically, for the defective portion 40 in FIG. 9A, in step S11, the shape of the defective portion is extracted, and a change in the shape of the color filter is confirmed. This shape is extracted by comparing the data of FIG. 9A with the shape of the basic pattern, and comparing the data of FIG.
The degree of change in the data shape is detected. As an example, when the data is out of the frame of ± 10% of the design data as the basic pattern, that is, when the line width and the area deviate from ± 10% of the design data, in step S12, the original data is obtained. When judging whether or not it is outside the color filter portion, it is judged that it is outside the color filter portion, and it is judged in Step S13 that the color filter is abnormal.

In step S14, it is determined whether or not the color of the defective portion, in which the shape of the defective portion is determined not to protrude from the original color filter in step S12, is different from the color of the original color filter. If the color is not different from the original color filter color, it is determined in step S15 that the pixel is defective. If the color is different from the color of the original color filter, it is determined that there is a foreign matter, not a pixel defect, and it is determined whether or not the foreign matter is on the panel surface. That is, in step S16, the focus position adjusting mechanism 75 of the magnifying intermediate lens 5 is operated as shown in FIG. 9B to detect the height position where the foreign matter is focused. Various methods can be considered as a focusing method. The simplest method is a method of setting a position where the density contrast of the foreign substance portion has the largest value as the focal point position. FIG.
As shown in (f), the sum of squares of the difference values of the pixel density data of a certain size is obtained as a focus evaluation value in the certain processing area 48.

[Number 2] focus evaluation value _{= Σ│g x + size, y -g} x, y │ 2 + Σ│
_{g x, y + size -g x} , y │ 2 where, g _{x, y} is the concentration data of the coordinates (x, y).

As shown in FIG. 9 (g), the height is changed within a certain range, and each focus evaluation value is obtained. Therefore,
In FIG. 9G, the position 47 having the largest focus evaluation value is set as the focal point. Note that the vertical axis in FIG. 9G indicates the focus evaluation value. In step S16, the focus position of the defects 41 and 42 is obtained by focusing on the defect portion. As a result, in step S17, it is determined whether the focus position is above the panel. As a result of this determination, it is detected that the defect 41 above the panel exists inside the liquid crystal panel 1 (step S19), while the defect 42 not above the panel exists on the surface of the liquid crystal panel 1. Is detected (step S18). As described above, since the surface foreign matter can be removed from the screen, it is desired to distinguish the foreign matter from the internal foreign matter, so that this processing is effective.

In the second embodiment, the magnifying intermediate lens 5 is used as a camera for performing the above-described processing. However, any camera may be used as long as it can perform color identification. Further, a spectral filter may be used in front of the monochrome camera. If color identification is unnecessary, this is not always the case. According to the second embodiment, in the image obtained by the magnifying intermediate lens 5, the screen is determined from the color, shape, and height of the defective portion detected by the camera 2 and whose detailed position is specified by the magnifying intermediate lens 5. It is possible to distinguish whether the defect is a pixel defect of the image itself, a defect mixed with a foreign substance on the screen itself, or a foreign substance adhered to the screen surface.

[0028]

According to the above arrangement, the defect detection and the defect position identification are shared, so that the number of cameras and the number of pixels of the image sensor can be increased regardless of the high definition of the component for detecting the defect, for example, the liquid crystal panel. Defect detection sensitivity and defect position identification accuracy can be ensured without relying on it. In addition, the present invention secures the defect detection sensitivity by a defect approximate position specifying system including an optical path changing rotary plate in an optical axis between a component to be detected and a camera, and enlarges a part of the component. A lens for taking the enlarged image into the camera can be further arranged on the optical axis, and the defect detailed position can be specified in a state where the lens is arranged, so that the defect position specifying accuracy can be secured. In other words, by inserting the defect detail position specifying lens 5 as an intermediate lens into the optical path, the pixel resolution can be considerably increased and a margin can be provided, so that defect detection sensitivity and defect position specifying accuracy can be ensured. Becomes possible. In addition, in order to cope with a large screen and high definition of a liquid crystal panel as an example of a component for detecting a defect, the defect detection sensitivity and the defect position identification accuracy are not dependent on the increase in the number of cameras and the number of pixels of an image sensor. It is possible to secure. Also, by using a lens, it is possible to obtain an enlarged image of a predetermined part of the component, and it is possible to distinguish various types of defects.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a configuration of a liquid crystal screen defect detection device according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating a flow of a defect detection process according to the first embodiment.

FIG. 3A is a diagram showing one state of an optical path changing rotary plate between a camera and a liquid crystal panel of the defect detection device;
(B) is a diagram showing an optical path shift on a pixel of an image sensor of the camera in the state of (a), and (c) is one of optical path changing rotating plates between the camera and the liquid crystal panel of the defect detection device. FIG. 6D is a diagram showing two states, and FIG. 6D is a diagram showing an optical path shift on a pixel of an image sensor of the camera in the state of FIG.

FIG. 4 is an explanatory diagram for explaining an optical path shift in the defect detection device.

FIGS. 5A and 5B are schematic diagrams each showing a light receiving position on an image sensor of a camera of the defect detection device. FIGS.

FIGS. 6 (a), (b), (c), and (d) are views showing images captured by a camera of the defect detection device.

FIGS. 7A and 7B are diagrams respectively showing images captured by a camera of the defect detection apparatus.

FIG. 8 is a flowchart illustrating a flow of a defect detection process of the defect detection device according to the second embodiment of the present invention.

9A is a diagram showing images of three defective portions, FIG.
(B) is a diagram of the device in a state where the image of (a) is captured,
(C) is a diagram showing a density profile of the confirmation line 45 of the defect 43, (d) is a diagram showing a density profile of the confirmation line 44 of the defect 41, and (e) is a confirmation line 4 of the defect 42.
6 is a diagram illustrating a density profile of FIG. 6, (f) is a diagram for explaining a focus evaluation value, and (g) is a diagram illustrating a focus evaluation value.

FIGS. 10A and 10B are schematic diagrams each showing a light receiving position on an image sensor of a camera of a conventional defect detection device.

[Explanation of symbols]

S1: LCD panel lighting S2: Camera image capture S3: Defect approximate position identification S4: Enlarged intermediate lens horizontal movement / focusing / image capture S5: Defect detail position identification S10: Color extraction of defect part S11: Shape of defect part Extraction S12: Shape determination S13: Color filter abnormality S14: Color determination S15: Pixel defect S16: Focus on defective part S17: Focus position determination S18: Foreign matter on panel surface S19: Foreign matter inside panel 1: Liquid crystal panel 2: Camera, 4: Computer, 5:
Enlarged intermediate lens, 6: optical path changing rotary plate, 7, 8: state of optical path changing rotary plate, 9: band pass filter, 20: captured image, 21: liquid crystal panel, 22, 23: pixel defect, 2
4, 25: pixel defect center pixel (maximum density), 40, 4
1, 42, 43: defect, 44, 45, 46: density profile confirmation line, 47: focus position, 48: processing area, 50: image sensor (pixel), 51: pixel defect of liquid crystal panel (on the image sensor) Received light), 52, 53, 5
4: pixel defect of the liquid crystal panel displaced by the rotating plate (light received on the image sensor); 75: focus position adjusting mechanism; 76: moving mechanism; 77: drive motor.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI H01J 9/42 H04N 5/66 102Z H04N 5/66 102 7/18 B 7/18 G06F 15/62 400

Claims (8)

[Claims] An image of a part (1) is captured by an image sensor (50), and a defect approximate position specifying camera (2) for roughly specifying a defect position of a point defect in the captured image; It is arranged rotatably in the optical axis between the image sensor and the component, and continuously rotates the relative optical axis position between the component and the image sensor by rotating around the rotation axis during imaging. An optical path changing rotary plate (6) capable of shifting the image of the component whose optical axis position is relatively displaced into the image sensor of the camera, and an image obtained by partially enlarging and expanding the component. A lens (5) for specifying a detailed position of a defect which is taken in by a camera; a general position of a point defect of the part is specified by the camera; Each pixel of the grayscale image A point defect detecting unit that detects the point defect from the density data and specifies a defect detailed position. 2. The point defect detection device according to claim 1, wherein the component is a liquid crystal panel, and inspects a point defect in a liquid crystal screen of the liquid crystal panel. 3. In an image obtained by the camera through the lens, from the color, shape, and height, it is determined whether the image is a pixel defect on the screen itself, a defect mixed in with the foreign substance on the screen itself, or a foreign substance on the screen surface. 3. The method according to claim 2, wherein the determination is made as to whether or not the adhesion is present.
3. The point defect detection device according to item 1. 4. A moving mechanism (76) for moving the lens in a direction orthogonal to an optical axis of the component,
4. The point defect detection apparatus according to claim 1, wherein the lens is moved in accordance with the information on the approximate position of the defect, the image of the defect is captured, and the detailed position of the defect is specified. 5. During imaging, the optical path changing rotary plate (6) is arranged in the optical axis between the image pickup device (50) of the camera (2) for identifying a defect approximate position and the component (1) to rotate the optical path changing rotary plate (6). By rotating about the axis, the relative optical axis position between the image sensor and the component is continuously shifted, and the image of the component whose optical axis position is relatively shifted is taken into the image sensor of the camera. Based on the image captured by the image sensor, the defect approximate position of a point defect in the image is specified, and the part is partially enlarged based on the defect approximate position of the specified point defect. An image of a part of the component is enlarged and captured by the camera through a lens (5) capable of capturing the enlarged image by the camera, and the density of each pixel of a grayscale image of the component captured by the camera through the lens. Detect the above point defects from the data A point defect detection method characterized in that a defect detailed position is specified by using the method. 6. The point defect detection method according to claim 5, wherein the component is a liquid crystal panel, and a point defect in a liquid crystal screen of the liquid crystal panel is inspected. 7. An image obtained by the camera through the lens, based on the color, shape and height, whether the image is a pixel defect on the screen itself, a foreign matter defect on the screen itself,
7. The point defect detection method according to claim 6, wherein it is determined whether or not foreign matter is attached to the screen surface. 8. A moving mechanism (76) for moving the lens in a direction orthogonal to an optical axis with the component,
8. The point defect detecting method according to claim 5, wherein the lens is moved in accordance with the information on the approximate position of the defect to capture an image of the defect to specify the detailed position of the defect.