TW201543023A - Bright spot inspection apparatus for filtering out foreign objects and noises and method thereof - Google Patents

Abstract

A bright spot inspection apparatus for filtering out foreign objects and noises comprises an inspection platform, an image capturing device, a back light source, a lateral light source, and an arithmetic unit. The inspection platform is for placing a to-be-inspected panel. The image capturing device is arranged at one side of the inspection platform for capturing surface images of the to-be-inspected panel according to an imaging command. The back light source is arranged at another side of the inspection platform opposing to the image capturing device. The lateral light source is arranged at a periphery side of a front surface and/or a back surface of the to-be-inspected panel. The arithmetic unit activates the back light source for obtaining a mixed defect surface image from the image capturing device and activates the lateral light source for obtaining a foreign object defect surface image from the image capturing device, and filters out the foreign object and noises on the mixed defect surface image according to the foreign object defect surface image so as to output a bright light distribution image.

Description

Bright spot detecting device for filtering foreign matter noise and method thereof

The invention relates to a bright spot detecting device, in particular to a bright spot detecting device capable of filtering foreign matter noise.

Currently in the industry, a complete panel process must pass at least the following tests: 1. Detection of glass substrates and color filters. 2. Panels, bright spots on the polarization sheet, and broken spot detection. 3. Detection of unevenness (Mura) is usually carried out in liquid crystal cell engineering or module engineering. 4. 瑕疵 detection in array circuit engineering.

Among them, the bright spots and broken bright spots on the panel and the polarizer are common defects in the panel manufacturing process. When the polarizer or panel is completed, the surface of the polarizer or panel is first inspected by the Automated Optical Inspection (AOI). Generally speaking, the surface state of the finished product is obtained by using an optical instrument, and the foreign matter or pattern abnormality is detected by computer image processing technology. Such bright spots and broken highlight images will be automatically classified when they are detected and marked as points to be tested for subsequent manual quality control (FQC) manual visual inspection. Make the final decision.

There are some techniques for panel detection. For such bright spots and broken bright spots, there are fixed detection steps. The detection steps are as follows: the panel to be tested is transported to the detection platform, and the upper and lower light sources (ie, backlights) are provided. The light is displayed on the panel to display a polarizing plate or a defect on the panel, and the displayed defect is obtained by shooting with a CCD camera. After the obtained image is binarized by the image processor, corresponding bright spots and broken bright points can be obtained. By further defining the aforementioned bright spots, and the position and coordinates of the broken bright spots, the enamel panels (bright spots, broken bright spots) can be repaired by laser welding to improve the yield of the product.

However, with the sputum image produced by the upper and lower light sources, in addition to the bright spots and broken bright spots generated in the process, it is possible to display foreign matter (such as dust and hair) attached to the panel and the polarizer. Foreign bodies of the class will cause an overkill in the detection process.

It is an object of the present invention to solve the problem of the occurrence of an overkill in a conventional detection method in which panel defects and foreign matter cannot be resolved.

In order to achieve the above object, the present invention provides a bright spot detecting device for filtering foreign matter noise, which is used for detecting a surface of a panel to be tested to obtain a bright spot distribution image of the panel to be tested, and the bright spot detecting device includes a a detecting platform provided on the measuring panel, an image capturing device disposed on one side of the detecting platform, a back light source disposed on a side of the detecting platform opposite to the image capturing device, and one or a plurality of panels disposed on the panel to be tested A lateral light source on the front side and/or a back side of the back side, and an arithmetic unit. The image capturing device takes the surface of the panel to be tested according to the image capturing instruction image. The operator includes an image capture indicator and an image processor. The image capturing indicator obtains a foreign object by activating the back light source and transmitting the image capturing command to the image capturing device to obtain a mixed surface image by activating the lateral light source and transmitting the image capturing command to the image capturing device. The surface image is obtained by filtering the foreign matter noise on the surface image of the mixed image according to the surface image of the foreign matter, thereby outputting a bright spot distribution image.

Further, the lateral light sources are respectively disposed on the two sides of the front side of the detecting platform and the two sides of the back side of the detecting platform.

Further, an angle between an illumination direction of the lateral light source and a surface of the panel to be tested is between 20 degrees and 30 degrees.

Further, an angle between an irradiation direction of the lateral light source and a surface of the panel to be tested is about 25 degrees.

Another object of the present invention is to provide a method for detecting a broken bright spot for filtering foreign matter noise, which is obtained by detecting the surface of the panel to be tested by the image capturing device to obtain a bright spot distribution image of the panel to be tested, and the method includes the following Step: providing a back light source to the panel to be tested, and taking the panel to be tested by the image capturing device to obtain a mixed surface image; providing one or more lateral light sources on the front side of the panel to be tested and/or providing a Or a plurality of lateral light sources on the back surface of the panel to be tested, and the panel to be tested is taken by the image capturing device to obtain a foreign matter surface image; and the foreign matter on the surface image of the mixed layer is filtered according to the surface image of the foreign matter Noise, by which a bright spot distribution image is output.

Further, the lateral light sources are respectively disposed on the detection platform The two sides of the front side and the two sides of the back side of the detecting platform.

Further, an angle between an illumination direction of the lateral light source and a surface of the panel to be tested is between 20 degrees and 30 degrees.

Further, an angle between an irradiation direction of the lateral light source and a surface of the panel to be tested is about 25 degrees.

Therefore, several embodiments of the present invention have the following beneficial technical effects over the prior art:

1. The invention can effectively filter out the foreign matter components in the panel image, thereby obtaining a relatively accurate bright spot distribution image.

2. When the angle between the illumination direction of the lateral light source and the surface of the panel to be tested is about 25 degrees, the false detection of the bright spot in the surface image of the foreign matter can be reduced, thereby improving the filtering of the foreign matter noise. rate.

100‧‧‧ Highlight detection equipment

P‧‧‧Down panel

10‧‧‧Testing platform

11A‧‧‧Detection platform front

11B‧‧‧Detector platform back

20‧‧‧Image capture device

30‧‧‧Back light source

31‧‧‧Lighting

32‧‧‧Diffuser board

40‧‧‧Side light source

40A‧‧‧Upside lateral light source

40B‧‧‧Lower lateral light source

50‧‧‧Operator

51‧‧‧Image indicator

52‧‧‧Image Processor

G1‧‧‧mixed surface image

G2‧‧‧ foreign object surface image

G3‧‧‧ Highlight distribution image

θ ₁ ‧‧‧lateral light source angle

θ ₂ ‧‧‧lateral light source angle

S201-S203‧‧‧Steps

S301-S305‧‧‧Steps

FIG. 1 is a block diagram of a bright spot detecting device of the present invention.

2 is a side view showing the bright spot detecting device of the present invention.

Figure 3 shows a screenshot of the 瑕疵 image.

FIG. 4 is a schematic flow chart of a bright spot detecting method according to the present invention.

Figure 5 shows the calculation flow of the bright spot distribution image.

The detailed description and technical content of the present invention are now in conjunction with the schema. described as follows. In addition, the drawings are not intended to limit the scope of the present invention, and the proportions thereof are not intended to limit the scope of the present invention.

The invention provides a bright spot detecting device for filtering foreign matter noise, wherein the bright spot detecting device is used for performing high-precision optical instrument detection, and uses machine vision (detecting lens) as a standard for detecting, thereby obtaining a panel. Or the position of the foreign object or pattern on the polarizer.

Please refer to FIG. 1 and FIG. 2 , which are schematic side views and block diagrams of the bright spot detecting device of the present invention. As shown in the figure, the present invention provides a bright spot detecting device 100 for filtering foreign matter noise for treating The surface of the measuring panel P is detected to obtain a bright spot distribution image of the panel P to be tested. The illuminating device 100 includes a detecting platform 10 for the panel P to be tested, an image capturing device 20 disposed on one side of the detecting platform 10, and a detecting platform 10 disposed on the detecting device 10 a side back light source 30, one or a plurality of side light sources 40 (40A, 40B) disposed on the front side of the detecting platform front surface 11A and/or the detecting platform back surface 11B, and a connection to the image capturing device 20 and the back light source 30 The operator 50 of the lateral light source 40.

The detecting platform 10 is configured to be disposed on the panel to be tested P. In a preferred embodiment, the detecting platform 10 can be a panel positioning bracket for positioning the panel, and the X-axis positioning is adjustable. The rod and the Y-axis positioning rod position the panel to be tested P on a fixed plane. In another embodiment, the detecting platform 10 can be a vacuum type adsorption device, and the panel P to be tested is adsorbed by a vacuum adsorption disk. Thereby, the panel to be tested P is positioned on a fixed plane. In addition, the detection platform 10 can also be a transport belt that carries the panel P to be tested, and the surface of the panel P to be tested is imaged when the panel P to be tested passes through the image capturing device 20. However, the above description is merely illustrative of several preferred embodiments of the present invention, and is not intended to be limited to the embodiments described above.

The image capturing device 20 is disposed on one side of the detecting platform 10 (shown above the detecting platform 10 in the drawing) for correspondingly capturing the panel P to be tested, thereby obtaining the panel to be tested. The surface image of P. The image capturing device 20 can be a CMOS line scan camera, a CMOS area scan camera, or a CCD (Charge Coupled Device) line scan camera, and the panel P can be tested. Image acquisition and automatic identification, thereby obtaining a 瑕疵 distribution image of the panel P to be tested.

The back light source 30 is disposed on the other side of the detecting platform 10 relative to the image capturing device 20 (corresponding to the lower side of the detecting platform 10 in the drawing). The back light source 30 can uniformly and uniformly illuminate the light source by the lamp 31 and the diffuser 32 disposed on the side of the lamp 31, whereby the panel and the polarizer (the panel P to be tested) can be uniformly illuminated. By the backlight of the back light source, bright spots, broken bright spots and foreign matter on the panel and the polarizer (the panel P to be tested) are displayed along with the illumination of the back light source 30. At this time, the image capturing device 20 will capture the mixed highlights.混 and foreign matter reflection (foreign noise) mixed 瑕疵 surface image G1 (as shown in Figure 3).

The lateral light source 40 is disposed on a front side of the front and/or back of the detection platform. In this embodiment, the lateral light source 40 can be divided into two parts: an upper lateral light source 40A and a lower lateral light source 40B (in the figure, corresponding to the detection flat Above and below the table 10). The front side light source 40A and the lower side light source 40B are polished on the front and back sides of the panel P to be tested, and the light irradiated onto the panel P to be tested is mostly unable to penetrate the panel due to the total reflection effect. Therefore, the enamel (bright spot, broken bright spot) on the panel and the polarizer will not be displayed on the panel P to be tested, and the foreign matter on the panel and the polarizer will be displayed by the lateral light irradiation. At this time, the image capturing device 20 captures the foreign matter surface image G2 including only the foreign matter reflection. In order to reduce the light transmittance of the lateral light source 40, the deposition rate of the bright spot and the broken bright spot is reduced. In a preferred embodiment, the angle between the illumination direction of the lateral light source 40 and the surface of the panel P to be tested is compared. The best is between 20 degrees and 30 degrees. This part will be supported by detailed experimental data in the latter part.

The computing unit 50 can be combined with the storage unit to form a computer or a processor, such as a personal computer, a workstation, a host computer, a mobile device, a tablet or other type of computer or processor, and the type is not limited herein. In this embodiment, the computing unit is, for example, a central processing unit (CPU), or other programmable general purpose or special purpose microprocessor (Microprocessor), digital signal processor (Digital Signal Processor) , DSP), programmable controller, Application Specific Integrated Circuits (ASIC), Programmable Logic Device (PLD) or other similar devices or a combination of these devices.

In the present invention, the computing unit 50 mainly includes an image capturing indicator 51 and an image processor 52. The image capturing indicator 51 is used to issue a start command and an image finger to the image capturing device 20, the back light source 30, and the side light source 40. Thereby, a mixed 瑕疵 surface image G1 and a foreign matter surface image G2 can be respectively obtained by the image capturing device 20. The image processor 52 can filter out the foreign matter noise component on the mixed 瑕疵 surface image G1 according to the foreign matter surface image G2, thereby outputting a bright spot distribution image G3 (as shown in FIG. 3).

The following is a detailed description of the detailed operation flow and corresponding calculation flow of the present invention: Please refer to "Fig. 3" and "Fig. 4" together, which is a schematic diagram showing a screenshot of the image and a flow chart of the bright spot detection method of the present invention, such as As shown in the figure, in the present invention, the bright spot detecting method is performed through the following steps. At the beginning of the detection process, first, the image capturing indicator 51 activates the back light source 30, and transmits an imaging command to the image capturing device 20 to capture and obtain the mixed 瑕疵 surface image G1 (step S201). In the continuation, the image capturing indicator 51 activates the side light source 40 and transmits an imaging command to the image capturing device 20 to capture and obtain the foreign matter surface image G2 (step S202). Finally, the image processor 52 filters out the foreign matter noise component on the mixed pupil surface image G1 according to the foreign matter surface image G2, thereby outputting a bright spot distribution image G3 (step S203).

The bright spot distribution image G3 can be obtained by the following calculation method, please refer to "Fig. 5". When acquiring the mixed 瑕疵 surface image G1 and the foreign matter 瑕疵 surface image G2, the image processor 52 performs gray scale processing and binarization processing on the mixed 瑕疵 surface image G1 and the foreign matter 瑕疵 surface image G2, respectively, to obtain an image. Feature points (step S301), and highlighting the feature points by coloring or marking to record coordinates and edge features of the feature points (step S302). Continuing, the corresponding feature points of the mixed 瑕疵 surface image G1 and the foreign matter 影像 surface image G2 are found by coordinates of the feature points and aligned for the edge features (step S303), and the mixed 瑕疵 surface image is filtered out A feature point corresponding to the foreign matter surface image G2 in G1 (step S304), and marking or coloring the remaining feature points (step S305), thereby obtaining the bright spot distribution image G3.

Through the above steps, the present invention obtained the following experimental data through the results of a plurality of experiments. The lateral light source angle θ ₁ in the experimental condition is an angle between the irradiation direction of the upper lateral light source 40A and the surface of the panel P to be tested, and the lateral light source angle θ ₂ is the lower side. An angle between the direction of illumination of the light source 40B and the surface of the panel P to be tested. The backlight detection refers to the number A of feature points that can be obtained in the mixed 瑕疵 surface image G1 when illuminated by the back light source 30, and the filtered detection image refers to the bright spot distribution image G3 after filtering. The number of remaining feature points B. The filtration rate is the percentage of the amount that can be filtered out P = {(AB) / A} x 100%.

As a result of the experiment, the following conclusion can be obtained: when the angle between the irradiation direction of the upper lateral light source and the lower lateral light source and the surface of the panel to be tested is horizontal (ie, θ ₁ = θ ₂ =0°), The filtration rate of foreign matter noise was 82.2%. When the angle between the illumination direction of the upper lateral light source and the surface of the panel to be tested is 25°, and the angle between the illumination direction of the lower lateral light source and the surface of the panel to be tested is horizontal (ie, when θ ₂ =0°), the foreign matter The noise filtering rate is 85%. When the angle between the upper lateral light source and the lower lateral light source and the surface of the panel to be tested is 25°, the foreign matter noise filtering rate can be increased to 90.5%.

The following experimental data was tested for commercial samples 1, commercial samples 2 and commercial samples 3 of different brand panels. In this test, the angle between the direction of the upper and lower lateral light sources and the surface of the panel to be tested is 25 degrees. As a control factor in the experiment, samples 1 and 2 are sold for different brands. 3 Tests were performed separately as operational variables in the experiment. In the experimental results, the backlight detection refers to the number A of feature points that can be obtained in the surface image of the mixed crucible when irradiated by the back light source, and the detected point is filtered after the filtering process. , the number B of remaining feature points. The filtration rate is the percentage of the amount that can be filtered out P = {(A - B) / A} x 100%.

According to the experimental results, the following conclusions can be drawn: in the test of the commercially available sample 1, the number of feature points that can be obtained by the backlight detection is 1484, and the remaining feature points in the processed bright spot distribution image are filtered out. The number is 234 and the filtration rate is 84.2%. In the test of the commercially available sample 2, the number of feature points that can be obtained by backlight detection is 1,718, and the number of feature points remaining in the image of the bright spot distribution after filtering is 190, and the filtering rate is 88.9. %. In the test of the commercially available sample 3, the number of feature points that can be obtained by backlight detection is 1500, and the number of feature points remaining in the filtered image of the bright spot after filtering is 168, and the filtering rate is 88.8. %. From the test results of the three samples, it can be seen that when the angle between the irradiation direction of the upper and lower lateral light sources and the surface of the panel to be tested is 25 degrees, the filtering rate of the foreign matter noise can reach 80% or more. . From the above experimental results, it can be seen that when the angle is approximately 25°, the detection rate of foreign matter noise can be greatly increased and false positives can be reduced.

In summary, the present invention can effectively filter out foreign matter components in the panel image, thereby obtaining a relatively accurate bright spot distribution image. In addition, when the angle between the illumination direction of the lateral light source of the present invention and the surface of the panel to be tested is about 25 degrees, the situation of erroneously detecting a bright spot in the surface image of the foreign matter can be reduced, thereby improving the filtering of the foreign matter noise. rate.

The present invention has been described in detail above, but the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. The equivalent changes and modifications made by the scope of the present invention should remain within the scope of the patent of the present invention.

100‧‧‧ Highlight detection equipment

P‧‧‧Down panel

10‧‧‧Testing platform

11A‧‧‧Detection platform front

11B‧‧‧Detector platform back

20‧‧‧Image capture device

30‧‧‧Back light source

31‧‧‧Lighting

32‧‧‧Diffuser board

40A‧‧‧Upside lateral light source

40B‧‧‧Lower lateral light source

θ ₁ ‧‧‧lateral light source angle

θ ₂ ‧‧‧lateral light source angle

Claims (8)

A bright spot detecting device for filtering foreign matter noise is used for detecting a surface of a panel to be tested to obtain a bright spot distribution image of the panel to be tested, the highlight detecting device comprising: a detecting platform for the panel to be tested The image capturing device is disposed on a side of the detecting platform, and the surface of the panel to be tested is imaged according to an image capturing instruction; a back light source is disposed on a side of the detecting platform opposite to the image capturing device; a plurality of lateral light sources disposed on a front side of the detecting platform and/or a peripheral side of a back surface; and an arithmetic unit including an image capturing indicator and an image processor, wherein the image capturing indicator is activated by The back light source transmits the image capturing command to the image capturing device to obtain a mixed surface image, and the image processor is obtained by activating the side light source and transmitting the image capturing command to the image capturing device to obtain a foreign matter surface image. The foreign matter noise on the surface image of the mixed image is filtered according to the surface image of the foreign matter, thereby outputting a bright spot distribution image. The bright spot detecting device of claim 1, wherein the lateral light sources are respectively disposed on the front side of the detecting platform and the back side of the detecting platform. The bright spot detecting device according to any one of the preceding claims, wherein the angle between the direction of illumination of the lateral light source and the surface of the panel to be tested is between 20 and 30 degrees. . The bright spot detecting device according to any one of claims 1 to 2, wherein an angle between an irradiation direction of the lateral light source and a surface of the panel to be tested is about 25 degrees. A method for detecting a bright spot for filtering foreign matter noise is to detect a surface of a panel to be tested by using an image capturing device to obtain a bright spot distribution image of the panel to be tested, the method comprising the steps of: providing a back light source to The panel to be tested, and the panel to be tested is taken by the image capturing device to obtain a mixed surface image; one or more lateral light sources are provided on the front surface of the panel to be tested and/or one or more lateral light sources are provided. And the surface of the panel to be tested is taken by the image capturing device to obtain a foreign matter surface image; and the foreign matter noise on the surface image of the mixed layer is filtered according to the surface image of the foreign matter, thereby outputting A bright spot distributes the image. The method for detecting a bright spot according to claim 5, wherein the lateral light sources are respectively disposed on the front side of the detecting platform and the back side of the detecting platform. The method for detecting a bright spot according to any one of the items 5 or 6, wherein the angle between the direction of the lateral light source and the surface of the panel to be tested is between 20 and 30 degrees. . The bright spot detecting method according to any one of the items 5 or 6, wherein the angle between the direction of illumination of the lateral light source and the surface of the panel to be tested is about 25 degrees.