An automatic detecting method of anisotropic conductive film particles in a liquid crystal screen

Abstract

The application provides an automatic detection method for anisotropic conductive film particles in a liquid crystal screen, comprising the following steps: S1, a picture acquisition step; S2, an image pre-positioning step; S3, a conductive particle region segmentation step; S4, a conductive particle detection step; and S5, a detection result derivation step. The application can automatically detect different models of liquid crystal screens, judge whether the number of conductive particles in the ACF is qualified, and has high detection efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and specifically to an automatic detection method for anisotropic conductive film particles in a liquid crystal display screen. Background Technology

[0002] Currently, LCD screens are becoming increasingly thinner and lighter, leading to ever-increasing demands on industrial production. Inspecting the connection between the LCD cell glass and IC devices is essential, primarily achieved using die-glass bonding (COG) technology. The core of COG technology is anisotropic conductive film (ACF). The conductivity of the LCD screen depends on the number of conductive particles in the ACF; therefore, the quality of the LCD screen's production can be judged based on the number of conductive particles.

[0003] After being pressed together, ACF particles exhibit distinct characteristics and a relatively uniform distribution, making them suitable for detection using an automated optical inspection (AOI) system. Compared to manual inspection, this method is faster and more accurate. An existing patent discloses an automatic detection method for pressed ACF conductive particles (Publication No.: CN 110672474 A), which uses a dynamic threshold segmentation algorithm to segment the conductive particle region. However, this algorithm performs poorly when there are multiple conductive particle regions within a single pin, and the detection of conductive particles is easily affected by impurities in the image, resulting in a high rate of missed detections and false detections. Summary of the Invention

[0004] Objective of the Invention: The objective of this invention is to provide an automatic detection method for anisotropic conductive film particles in a liquid crystal display (LCD) screen, thereby addressing the problems of poor segmentation performance and low accuracy of conductive particle detection algorithms in the aforementioned background technology. The method of this invention specifically includes the following steps:

[0005] Step 1: Acquire images of the conductive particle region;

[0006] Step 2: Locate the marker points in the image, then binarize the resulting image and calculate the coordinates of the center of the marker points;

[0007] Step 3: Use a mean square error-based algorithm to segment each conductive particle region;

[0008] Step 4: Use a local maximum algorithm to detect each region segmented in Step 3;

[0009] Step 5: Count the number of conductive particles in Step 4 and compare it with the preset threshold to determine whether the number of conductive particles in the anisotropic conductive film ACF is qualified.

[0010] In step 1, the DIC differential interferometer lens DAF-100 is used in conjunction with the DALSA linear array camera to acquire images, so that the conductive particle imaging has a three-dimensional effect and is easy to separate from the background image.

[0011] Step 2 includes:

[0012] Step 2-1: Perform grayscale processing on the image acquired in Step 1 using a weighted method. The calculation formula for each pixel is as follows:

[0013] Gray=0.299×R+0.587×G+0.114×B

[0014] Where Gray represents grayscale value, R represents red value, G represents green value, and B represents blue value;

[0015] Step 2-2: Using the cross-shaped markers as templates, the image is processed using a template matching algorithm to obtain matching images. The similarity calculation formula is as follows:

[0016]

[0017] Where D(i,j) represents the similarity at coordinates (i,j), M is the maximum value of the horizontal axis, N is the maximum value of the vertical axis, T(s,t) represents the energy of the template, and S(i+s-1,j+t-1) is the energy of the corresponding sub-image under the template coverage.

[0018] Steps 2-3: Use the perceptual hash algorithm to determine the similarity between the template image and the matched image. Compare the number of bits in the results. If the number of different bits is less than X1 (usually 10), the images are considered similar; otherwise, they are considered dissimilar.

[0019] Step 2-4: Use Otsu's method to binarize images that are similar to those determined in step 2-3;

[0020] Steps 2-5: Impurities generated after binarization are removed sequentially through erosion, area thresholding, and dilation operations.

[0021] Step 2-6: Extract the image of the cross-shaped markers from Step 2-5, and then use the centroid of the markers as the positioning point, with the centroid coordinates (x...). c ,y c The calculation formula for ) is as follows:

[0022]

[0023] Where T is the number of pixels in the cross-shaped marker image, X(t) is the horizontal coordinate sequence, and Y(t) is the vertical coordinate sequence.

[0024] Step 3 includes:

[0025] Step 3-1: Perform mean squared error processing on the image of the conductive particle region. Traverse each pixel in the image and calculate the mean squared error within the 5×5 neighborhood of each pixel using the following formula:

[0026]

[0027] Where g(i,j) is the new gray value of the pixel, f(x,y) is the gray value of the pixel in the neighborhood, p represents the average gray value in the 5×5 neighborhood of the pixel (i,j), and q is the neighborhood size.

[0028] Step 3-2: Binarize the image processed in Step 3-1 using the OTSU method;

[0029] Step 3-3: After image binarization, there is a significant difference in grayscale values ​​between the area where the conductive particles are located and the background area. Therefore, grayscale projection is performed in the vertical direction and a grayscale projection map is drawn.

[0030] Steps 3-4: The grayscale value of the gap between the conductive particle regions in the grayscale projection image is 0. This characteristic is used to segment the pins.

[0031] Step 4 includes:

[0032] Step 4-1: Enhance the image using a local mean difference algorithm based on illumination direction: Traverse each pixel in the image, divide the pixel into two parts within a 7×7 neighborhood, calculate the mean gray value of the pixels in the upper and lower parts respectively, and then calculate the difference between the mean gray values ​​of the pixels in the upper and lower parts. The absolute value of the difference is the new gray value of the target pixel, and finally obtain the mean difference image.

[0033] Step 4-2: Binarize the image processed in Step 4-1 using the Otsu method to obtain a binarized image;

[0034] Step 4-3: Use the breadth-first search algorithm to label the connected regions of the binarized image and calculate the area of ​​each connected region;

[0035] Step 4-4: Perform area threshold denoising: Set an area threshold Q. Since the size of conductive particles is generally between 7×7 and 11×11, Q is set to 49. Compare the area of ​​each connected region with the threshold Q. If it is less than Q, the connected region is judged as noise and the noise is deleted from the graph. Otherwise, the connected region is retained.

[0036] Step 4-5: Determine the position and number of conductive particles using the local maximum method: Traverse the connected regions of the mean difference image obtained in Step 4-1. Using the currently traversed pixel as the center and X2 (usually 5) as the radius, check if the pixel's gray value is a maximum. If it is, the pixel is the center of the conductive particle; otherwise, traverse to the next point.

[0037] In steps 4-5, if there are more than two maxima within a range, connect any two maxima and take the middle point as the new maxima, until there is only one maxima. This maxima is the center of the conductive particle.

[0038] The beneficial effects of this invention are as follows: It realizes the automatic detection of the position and number of conductive particles in ACF based on optical and image fusion method. Compared with manual sampling, this invention has a faster detection speed and higher accuracy, which can meet the needs of industrial testing. Attached Figure Description

[0039] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0040] Figure 1 This is a flowchart of the automatic detection of ACF particles in this invention.

[0041] Figure 2 A flowchart for an image pre-positioning algorithm.

[0042] Figure 3 Flowchart for implementing the perceptual hash algorithm.

[0043] Figure 4 This is the result of binarization.

[0044] Figure 5 This is a diagram with crosshair markers.

[0045] Figure 6 This is a diagram showing the positioning results.

[0046] Figure 7 This is a flowchart of the conductive particle region segmentation algorithm.

[0047] Figure 8 This is the mean squared error image.

[0048] Figure 9 This is the binarized grayscale projection image.

[0049] Figure 10 This is a flowchart of the conductive particle detection algorithm.

[0050] Figure 11 This is a schematic diagram of the algorithm for calculating the local mean difference in the direction of illumination.

[0051] Figure 12 This is a graph showing the detection results of conductive particles. Detailed Implementation

[0052] like Figure 1 As shown, this embodiment of the invention provides an automatic detection method for anisotropic conductive film particles in a liquid crystal screen, comprising the following steps:

[0053] S1. Image acquisition step: Use the DIC differential interferometer lens DAF-100 in conjunction with the DALSA linear array camera to acquire images, so that the conductive particle images have a three-dimensional effect and are easy to separate from the background image.

[0054] S2. Image pre-location step: Use template matching algorithm and perceptual hash algorithm to locate the marker points in the image, then perform binarization processing on the result image, and calculate and determine the coordinates of the center of the marker points.

[0055] S3, Conductive particle region segmentation step: Each conductive particle region is segmented using a mean square error-based algorithm.

[0056] S4, Conductive Particle Detection Step: Use a local maximum algorithm to detect each region segmented in step S3.

[0057] S5. Detection result export step: Count the number of conductive particles in step S4, compare it with the preset threshold, and determine whether the number of conductive particles in ACF is qualified.

[0058] In step S1, the ACF conductive particles are micrometers in size and the image is three-dimensional. Ordinary cameras cannot distinguish the protrusions of the conductive particles or the difference in grayscale values ​​between them and the background. Therefore, the invention uses a DAF-100 differential interferometry lens paired with a DALSA LA-CM-02K08A linear array camera for image acquisition.

[0059] like Figure 2 As shown, in step S2, template matching algorithm and perceptual hashing algorithm are used to locate the marker points in the image, and then the resulting image is binarized to calculate and determine the coordinates of the center of the marker points. The specific steps include:

[0060] S21. The image obtained in step S1 is converted to grayscale using a weighted method. The calculation formula for each pixel is as follows:

[0061] Gray=0.299×R+0.587×G+0.114×B

[0062] Where Gray represents grayscale value, R represents red value, G represents green value, and B represents blue value.

[0063] S22. Using the cross-shaped markers as templates, the image is processed using a template matching algorithm to obtain matching images. The similarity calculation formula is as follows:

[0064]

[0065] S23. Use the perceptual hash algorithm to determine the similarity between the template image and the matched image. Compare the number of bits in the results. If the number of different bits is less than 10, the images are similar; otherwise, the images are not similar. The specific algorithm implementation flowchart is as follows: Figure 3 As shown.

[0066] S24. For images with similar judgment results in steps 2-3, perform binarization using the Otsu method. The result is shown in the image below. Figure 4 As shown.

[0067] S25. After sequentially performing erosion, area thresholding, and dilation operations, the impurities generated after binarization are removed. The result is shown in the figure below. Figure 5 As shown.

[0068] S26. Following step S25, the image of the cross-shaped marker points can be extracted, and then the centroid of the marker points is used as the positioning point. Centroid coordinates (x...) c ,y c The calculation formula for ) is as follows:

[0069]

[0070] Where T is the number of pixels in the cross-shaped marker image, X(t) is the horizontal coordinate sequence, and Y(t) is the vertical coordinate sequence. The image of the localization result is as follows: Figure 6 As shown.

[0071] In step S2, the coordinates of the positioning point to the position of the conductive particle region are fixed. Therefore, once the positioning point is determined, the conductive particle region can be found subsequently.

[0072] like Figure 7 As shown, in step S3, an algorithm based on mean square error is used to segment each conductive particle region. The specific steps include:

[0073] S31. Perform mean square error processing on the image of the conductive particle region. Traverse each pixel in the image and calculate the error within the 5×5 neighborhood of that pixel using the following formula:

[0074]

[0075] Where g(i,j) is the new gray value of the pixel, f(x,y) is the gray value of the pixel in the neighborhood, p represents the average gray value in the 5×5 neighborhood of pixel (i,j), and q is the neighborhood size; the processed result is shown in the figure below. Figure 8 As shown

[0076] S32. Use the OTSU method to binarize the processed image.

[0077] S33. After image binarization, there is a significant difference in grayscale values ​​between the area where the conductive particles are located and the background area. Therefore, grayscale projection is performed in the vertical direction and a grayscale projection map is drawn.

[0078] S34. In the grayscale projection image, the grayscale value of the gap between the conductive particle regions is 0. This characteristic is used to segment the pins. The projection image is as follows: Figure 9 As shown.

[0079] like Figure 10 As shown, in step S4, an algorithm based on local maxima is used to detect conductive particles. The specific steps include:

[0080] S41. An algorithm using local mean difference in illumination direction is used to enhance the features of conductive particles. For each pixel in the image, iterates through the 7×7 neighborhood of that pixel, dividing it into upper and lower parts. The mean grayscale values ​​of the pixels in both parts are calculated, and the difference between them is calculated. The absolute value of this difference is the new grayscale value of the target pixel. The principle diagram of the algorithm is shown below. Figure 11 As shown.

[0081] S42. Then, the mean difference image is binarized using the OTSU method.

[0082] S43. Use the breadth-first search algorithm to label the connected regions of the binarized image and calculate the area of ​​each connected region.

[0083] S44. Next, area threshold denoising is performed. Specifically, an area threshold Q is set. Since the size of conductive particles is generally between 7×7 and 11×11, Q is set to 49. The area of ​​each connected region is compared with the threshold Q. If it is less than Q, the region is judged as noise and deleted from the graph. Otherwise, the region is retained.

[0084] S45. Finally, the location and number of conductive particles are determined using the local maximum method: The connected regions of the mean difference image obtained in step 4-1 are traversed. Using the currently traversed pixel as the center and a radius of 5, it is checked whether the grayscale value of that point is a maximum. If it is, then that point is the center of the conductive particle; otherwise, the next point is traversed. Note that if there are multiple maximum points within a range, any two maximum points are connected, and the middle point is taken as the new maximum point, until only one maximum point remains. This point is the center of the conductive particle. The detection result is shown in the image below. Figure 12 As shown.

[0085] In its specific implementation, this application provides a computer storage medium and a corresponding data processing unit. The computer storage medium is capable of storing a computer program, which, when executed by the data processing unit, can run the invention's content regarding an automatic detection method for anisotropic conductive film particles in a liquid crystal screen, as well as some or all of the steps in various embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0086] Those skilled in the art will clearly understand that the technical solutions in the embodiments of the present invention can be implemented using computer programs and their corresponding general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of computer programs, i.e., software products. These computer program software products can be stored in a storage medium and include several instructions to cause a device containing a data processing unit (which may be a personal computer, server, microcontroller, MUU, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0087] This invention provides an automatic detection method for anisotropic conductive film particles in a liquid crystal display screen. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment of the invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. An automatic detection method for anisotropic conductive film particles in a liquid crystal display screen, characterized in that, Includes the following steps: Step 1: Acquire images of the conductive particle region; Step 2: Locate the marker points in the image, then binarize the resulting image and calculate the coordinates of the center of the marker points; Step 3: Use a mean square error-based algorithm to segment each conductive particle region; Step 4: Use a local maximum algorithm to detect each region segmented in Step 3; Step 5: Count the number of conductive particles in Step 4 and compare it with the preset threshold to determine whether the number of conductive particles in the anisotropic conductive film ACF is qualified. Step 3 includes: Step 3-1: Perform mean squared error processing on the image of the conductive particle region. Traverse each pixel in the image and calculate the mean squared error within the 5×5 neighborhood of each pixel using the following formula: Where g(i,j) is the new gray value of the pixel, f(x,y) is the gray value of the pixel in the neighborhood, p represents the average gray value in the 5×5 neighborhood of the pixel (i,j), and q is the neighborhood size; Step 3-2: Binarize the image processed in Step 3-1 using the OTSU method; Step 3-3: After image binarization, there is a significant difference in grayscale values ​​between the area where the conductive particles are located and the background area. Therefore, grayscale projection is performed in the vertical direction and a grayscale projection map is drawn. Steps 3-4: The grayscale value of the gap between the conductive particle regions in the grayscale projection image is 0. This characteristic is used to segment the pins. Step 4 includes: Step 4-1: Enhance the image using a local mean difference algorithm based on illumination direction: Traverse each pixel in the image, divide the pixel into two parts within a 7×7 neighborhood, calculate the mean gray value of the pixels in the upper and lower parts respectively, and then calculate the difference between the mean gray values ​​of the pixels in the upper and lower parts. The absolute value of the difference is the new gray value of the target pixel, and finally obtain the mean difference image. Step 4-2: Perform binarization on the image processed in Step 4-1 using Otsu's method to obtain a binarized image; Step 4-3: Use the breadth-first search algorithm to label the connected regions of the binarized image and calculate the area of ​​each connected region; Step 4-4: Perform area threshold denoising: Set an area threshold Q, compare the area of ​​each connected region with the threshold Q. If the area is less than Q, the connected region is determined to be noise and the noise is deleted from the graph; otherwise, the connected region is retained. Step 4-5: Determine the position and number of conductive particles using the local maximum method: Traverse the connected regions of the mean difference image obtained in Step 4-1, and check whether the gray value of the pixel is a maximum value with the current pixel as the center and X2 as the radius. If it is, the pixel is the center of the conductive particle; otherwise, traverse the next point. In steps 4-5, if there are more than two maxima within a range, connect any two maxima and take the middle point as the new maxima, until there is only one maxima. This maxima is the center of the conductive particle.

2. The method according to claim 1, characterized in that, In step 1, the DIC differential interferometer lens DAF-100 is used in conjunction with the DALSA linear array camera to acquire images.

3. The method according to claim 2, characterized in that, Step 2 includes: Step 2-1: Perform grayscale processing on the image acquired in Step 1 using a weighted method. The calculation formula for each pixel is as follows: Gray=0.299×R+0.587×G+0.114×B Where Gray represents grayscale value, R represents red value, G represents green value, and B represents blue value; Step 2-2: Using the cross-shaped markers as templates, the image is processed using a template matching algorithm to obtain matching images. The similarity calculation formula is as follows: Where D(i,j) represents the similarity at coordinate (i,j), M is the maximum value of the horizontal axis, N is the maximum value of the vertical axis, T(s,t) represents the energy of the template, and S(i+s-1,j+t-1) is the energy of the corresponding sub-image under the template coverage. Steps 2-3: Use the perceptual hash algorithm to determine the similarity between the template image and the matched image. Compare the number of bits in the results. If the number of different bits is less than X1, the images are considered similar; otherwise, the images are considered dissimilar. Step 2-4: Use Otsu's method to binarize images that are similar to those determined in step 2-3; Steps 2-5: Impurities generated after binarization are removed sequentially through erosion, area thresholding, and dilation operations. Step 2-6: Extract the image of the cross-shaped markers from Step 2-5, and then use the centroid of the markers as the positioning point, with the centroid coordinates (x...). c ,y c The calculation formula for ) is as follows: Where T is the number of pixels in the cross-shaped marker image, X(t) is the horizontal coordinate sequence, and Y(t) is the vertical coordinate sequence.

Images