Method and device for automatically detecting number and deflection of conductive particles of liquid crystal screen

Abstract

The application discloses a kind of liquid crystal screen conductive particle number and deflection automatic detection method, comprising the following steps: S1: using optical system to the region to be detected is imaged, obtains the conductive particle image to be detected;S2;Conductive particle image to be detected is carried out image enhancement processing, improve the contrast of image, highlight conductive particle;S3: the conductive particle image after image enhancement processing is segmented, obtains bump region set;S4: the bump region set obtained calculates the number of conductive particles;S5: the conductive particle mapped to the corresponding blank bump projection region is calculated, and the deflection of conductive particle is calculated.Correspondingly, the application also provides a kind of liquid crystal screen conductive particle number and deflection automatic detection device.The present application solves the problem of low precision of liquid crystal screen conductive particle counting and low precision of deflection detection of conductive particle.

Description

Technical Field

[0001] This invention belongs to the fields of automation and computer vision technology, and specifically relates to an automatic detection method and device for the number and misalignment of conductive particles in a liquid crystal display screen. Background Technology

[0002] In an LCD screen, besides the large LCD panel itself, a driver chip must be connected to its periphery to control the display signals. Currently, COG or FOG technologies are commonly used to bond flexible printed circuit boards (FPCs) to the LCD glass. The core of the bonding process is the anisotropic conductive film (ACF). ACF mainly consists of conductive particles and thermosetting resin. During bonding, high temperature and pressure are applied to connect the electrodes on the COF side and the electrodes on the FPC side through the conductive particles, achieving circuit conductivity. In practice, bonding defects may occur, such as insufficient particle count or poor bonding between the COF and FPC electrodes, leading to poor panel performance and function. Therefore, the bonding effect is usually inspected after the bonding process. Due to the low light transmittance of the bonding area, although the ACF particles have certain characteristics after bonding, ordinary optical systems cannot effectively image the bonded state of the conductive particles. Specific optical systems are required for imaging. Automated detection of the image using algorithms can greatly improve detection speed and accuracy.

[0003] Chinese patent publication number "CN107144602A" discloses a method for detecting a circuit consisting of a detection point, an electrode, and a connector using physical methods. This method requires reserving a detection area at the electrode position on the printed circuit board, which is complex to operate and not conducive to achieving automatic detection.

[0004] Chinese patent publication number CN110672474A discloses an automatic detection method for ACF conductive particle bonding, including the following steps: S1, image acquisition; S2, primary conductive particle region detection; S3, secondary conductive particle region detection; S4, particle region selection; S5, particle region verification; S6, conductive particle judgment; S7, bonding result export. This invention does not statistically analyze particle misalignment. Furthermore, due to the low light transmittance of the bonding area material, the contrast between the imaged particles and the background is generally low, and the detection area often contains significant interference from impurities. This invention directly segments the image according to a threshold, resulting in low detection accuracy, which cannot meet the current requirements for high-precision particle counting and misalignment detection. Summary of the Invention

[0005] This invention provides an automatic detection method and apparatus for the number and misalignment of conductive particles in a liquid crystal display (LCD) screen, aiming to solve the problems of low accuracy in counting conductive particles and low accuracy in detecting the misalignment of conductive particles in current LCD screens.

[0006] To address the aforementioned technical problems, this invention proposes an automatic detection method for the number and offset of conductive particles in a liquid crystal display (LCD) screen, comprising the following steps: S1: Use an optical system to image the area to be detected and obtain an image of the conductive particles to be detected.

[0007] S2; Perform image enhancement processing on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles.

[0008] S3: Segment the conductive particle image after image enhancement to obtain a set of bump regions.

[0009] S4: Calculate the number of conductive particles in the obtained bump region set.

[0010] S5: Map the calculated conductive particles to the corresponding blank bump projection area and calculate the displacement of the conductive particles.

[0011] S6: Output the number of conductive particles and the result of the offset.

[0012] Preferably, the image enhancement processing of the conductive particle image to be detected in step S2 specifically includes: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected, eliminate noise and smooth the image.

[0013] S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically:

[0014] In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location.

[0015] S23: Set the threshold constant for grayscale difference, if the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows:

[0016]

[0017] In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale values ​​of pixels used to enhance an image.

[0018] S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically:

[0019] In the formula, A low threshold for the grayscale value of pixels used to enhance the image.

[0020] S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically:

[0021] In the formula, This refers to the grayscale value of the pixel after image enhancement.

[0022] Preferably, obtaining the bump region set in step S3 specifically involves: S31: Project the conductive particle image along the Y-axis, calculate the sum of the gray values ​​of each row of pixels in the conductive particle image, and obtain the set of sums of gray values ​​of horizontal pixels.

[0023] S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories.

[0024] S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

[0025] Preferably, the calculation of the number of conductive particles in the obtained bump region set in step S4 specifically involves: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image; S42: Binarize the set of bump regions processed by Gaussian filtering to obtain a binary image; S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3.

[0026] S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated.

[0027] Preferably, the calculation of the displacement of the conductive particles in step S5 specifically involves: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections.

[0028] S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows:

[0029] In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the center coordinates of the rectangle corresponding to the bump area.

[0030] S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

[0031] Accordingly, the present invention also proposes an automatic detection device for the number and misalignment of conductive particles in a liquid crystal display screen. The device includes an optical system and electronic equipment. The electronic equipment includes a processor, a memory, and a computer program stored in the memory and executable on the processor. The device performs the following steps: S1: Use an optical system to image the area to be detected and obtain an image of the conductive particles to be detected.

[0032] S2; Perform image enhancement processing on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles.

[0033] S3: Segment the conductive particle image after image enhancement to obtain a set of bump regions.

[0034] S4: Calculate the number of conductive particles in the obtained bump region set.

[0035] S5: Map the calculated conductive particles to the corresponding blank bump projection area and calculate the displacement of the conductive particles.

[0036] S6: Output the number of conductive particles and the result of the offset.

[0037] Preferably, the image enhancement processing of the conductive particle image to be detected in step S2 specifically includes: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected, eliminate noise and smooth the image.

[0038] S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically:

[0039] In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location.

[0040] S23: Set the threshold constant for grayscale difference, if the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows:

[0041]

[0042] In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale values ​​of pixels used to enhance an image.

[0043] S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically:

[0044] In the formula, A low threshold for the grayscale value of pixels used to enhance the image.

[0045] S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically:

[0046] In the formula, This represents the grayscale value of a pixel after image enhancement.

[0047] Preferably, obtaining the bump region set in step S3 specifically involves: S31: Project the conductive particle image along the Y-axis, calculate the sum of the gray values ​​of each row of pixels in the conductive particle image, and obtain the set of sums of gray values ​​of horizontal pixels.

[0048] S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories.

[0049] S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

[0050] Preferably, the calculation of the number of conductive particles in the obtained bump region set in step S4 specifically involves: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image.

[0051] S42: Binarize the set of bump regions processed by Gaussian filtering to obtain a binary image.

[0052] S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3.

[0053] S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated.

[0054] Preferably, the calculation of the displacement of the conductive particles in step S5 specifically involves: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections.

[0055] S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows:

[0056] In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the center coordinates of the rectangle corresponding to the bump area.

[0057] S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

[0058] Compared with the prior art, the present invention has the following technical effects: 1. This invention proposes an automatic detection method and device for the number and misalignment of conductive particles in a liquid crystal display (LCD) screen. This invention can detect the number and misalignment of conductive particles with high accuracy required by current manufacturing processes. For a single bump area, when the number of conductive particles is ≤10, the detection error is less than 1; when the number of conductive particles is >10, the detection error is less than 10% of the number of particles. The detection accuracy for conductive particle misalignment is ±5μm, and the algorithm of this invention is simple and can meet the real-time requirements of industrial applications.

[0059] 2. This invention enhances the conductive particle image acquired by the optical system and divides it into bump and non-bump regions according to a calculated threshold. Before counting the number and offset of conductive particles in the bump region, a series of image processing operations are performed to remove image noise and interference, while preserving the main structure and boundaries in the image, thus effectively extracting the feature information of conductive particles. Attached Figure Description

[0060] Figure 1 This is an overall flowchart of the automatic detection method for the number and offset of conductive particles in a liquid crystal screen according to the present invention. Figure 2This is an imaging schematic diagram of the conductive particle image to be detected according to the present invention; Figure 3 This is a schematic diagram of the image of the enhanced conductive particles described in this invention; Figure 4 This is a schematic diagram illustrating the statistical counting of conductive particles as described in this invention; Figure 5 This is a schematic diagram of the conductive particle misalignment described in this invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present application and with reference to the accompanying drawings.

[0062] Example 1 See Figure 1 As shown, this invention proposes an automatic detection method for the number and offset of conductive particles in a liquid crystal display screen, comprising the following steps: S1: See Figure 2 As shown, an optical system is used to image the area to be detected, thereby acquiring an image of the conductive particles to be detected.

[0063] S2; see also Figure 3 As shown, image enhancement processing is performed on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles.

[0064] S3: Segment the conductive particle image after image enhancement to obtain a set of bump regions.

[0065] S4: See also Figure 4 As shown, the number of conductive particles is calculated for the obtained bump region set.

[0066] S5: See Figure 5 As shown, the calculated conductive particles are mapped onto the corresponding blank bump projection area, and the displacement of the conductive particles is calculated.

[0067] S6: Output the number of conductive particles and the result of the offset.

[0068] Furthermore, the image enhancement processing of the conductive particle image to be detected in step S2 specifically includes: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected. The convolution kernel of the Gaussian filter is generally set to 3 or 5. Use a 3*3 or 5*5 window to perform convolution operation on the image. Perform weighted calculation on the center pixel according to the set convolution kernel to eliminate noise and smooth the image.

[0069] S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically:

[0070] In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location.

[0071] S23: Set the threshold constant for grayscale difference. The threshold constant can be set according to actual needs. In this embodiment, the threshold constant is set to 140. If the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows:

[0072]

[0073] In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale values ​​of pixels used to enhance an image.

[0074] S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically:

[0075] In the formula, A low threshold for the grayscale value of pixels used to enhance the image.

[0076] S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically:

[0077] In the formula, This represents the grayscale value of a pixel after image enhancement.

[0078] Furthermore, the specific steps in step S3 for obtaining the bump region set are as follows: S31: Project the conductive particle image along the Y-axis, and automatically calculate the sum of gray values ​​of each row of pixels in the conductive particle image based on the projection result to obtain the set of sums of gray values ​​of horizontal pixels.

[0079] S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories.

[0080] S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

[0081] Furthermore, in step S4, calculating the number of conductive particles in the obtained bump region set specifically involves: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image.

[0082] S42: Binarize the set of bump regions after Gaussian filtering. Here, the Otsu method is used to automatically calculate and obtain a binary image. S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3.

[0083] S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated.

[0084] Furthermore, the calculation of the displacement of the conductive particles in step S5 specifically involves: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections.

[0085] S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows:

[0086] In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the center coordinates of the rectangle corresponding to the bump area.

[0087] S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

[0088] See Figure 5 As shown in the figure, the centroid of the conductive particles projected onto the selected rectangular frame of the corresponding bump region is pt2, and the center of the corresponding rectangular frame of the bump region is pt1. The displacement value of the conductive particles in the bump region can be obtained by calculating the distance difference between pt2 and pt1.

[0089] Example 2 Accordingly, this invention also proposes an automatic detection device for the number and misalignment of conductive particles in a liquid crystal display screen. The device includes an optical system and electronic equipment. The optical system employs a DIC differential interferometer lens paired with a DALSA line scan camera. The electronic equipment includes a processor, a memory, and a computer program stored in the memory and executable on the processor. The device performs the following steps: S1: See Figure 2 As shown, an optical system is used to image the area to be detected, thereby acquiring an image of the conductive particles to be detected.

[0090] S2; see also Figure 3 As shown, image enhancement processing is performed on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles.

[0091] S3: Segment the conductive particle image after image enhancement to obtain a set of bump regions.

[0092] S4: See also Figure 4 As shown, the number of conductive particles is calculated for the obtained bump region set.

[0093] S5: See Figure 5 As shown, the calculated conductive particles are mapped onto the corresponding blank bump projection area, and the displacement of the conductive particles is calculated.

[0094] S6: Output the number of conductive particles and the result of the offset.

[0095] Furthermore, the image enhancement processing of the conductive particle image to be detected in step S2 specifically includes: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected. The convolution kernel of the Gaussian filter is generally set to 3 or 5. Use a 3*3 or 5*5 window to perform convolution operation on the image. Perform weighted calculation on the center pixel according to the set convolution kernel to eliminate noise and smooth the image.

[0096] S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically:

[0097] In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location.

[0098] S23: Set the threshold constant for grayscale difference. The threshold constant can be set according to actual needs. In this embodiment, the threshold constant is set to 140. If the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows:

[0099]

[0100] In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale values ​​of pixels used to enhance an image.

[0101] S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically:

[0102] In the formula, A low threshold for the grayscale value of pixels used to enhance the image.

[0103] S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically:

[0104] In the formula, This refers to the grayscale value of the pixel after image enhancement.

[0105] Furthermore, the specific steps in step S3 for obtaining the bump region set are as follows: S31: Project the conductive particle image along the Y-axis, and automatically calculate the sum of gray values ​​of each row of pixels in the conductive particle image based on the projection result to obtain the set of sums of gray values ​​of horizontal pixels.

[0106] S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories.

[0107] S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

[0108] Furthermore, in step S4, calculating the number of conductive particles in the obtained bump region set specifically involves: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image.

[0109] S42: Binarize the set of bump regions after Gaussian filtering. Here, the Otsu method is used to automatically calculate and obtain a binary image. S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3.

[0110] S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated.

[0111] Furthermore, the calculation of the displacement of the conductive particles in step S5 specifically involves: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections.

[0112] S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows:

[0113] In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the center coordinates of the rectangle corresponding to the bump area.

[0114] S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

[0115] See Figure 5 As shown in the figure, the centroid of the conductive particles projected onto the selected rectangular frame of the corresponding bump region is pt2, and the center of the corresponding rectangular frame of the bump region is pt1. The offset value of the conductive particles in the bump region can be obtained by calculating the distance difference between pt2 and pt1.

[0116] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the inventive concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. An automatic detection method for the number and misalignment of conductive particles in a liquid crystal display screen, characterized in that, Includes the following steps: S1: Use an optical system to image the area to be detected and obtain an image of the conductive particles to be detected; S2; Perform image enhancement processing on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles; S3: Segment the image of conductive particles after image enhancement to obtain a set of bump regions; S4: Calculate the number of conductive particles in the obtained bump region set; S5: Map the calculated conductive particles to the corresponding blank bump projection area and calculate the displacement of the conductive particles; S6: Output the number of conductive particles and the result of the offset; The specific steps in step S4 for calculating the number of conductive particles in the obtained bump region set are as follows: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image; S42: Binarize the set of bump regions processed by Gaussian filtering to obtain a binary image; S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3. S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated. The calculation of the displacement of the conductive particles in step S5 is specifically as follows: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections. S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows: In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the coordinates of the center of the rectangle corresponding to the bump area; S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

2. The automatic detection method for the number and offset of conductive particles in a liquid crystal display screen according to claim 1, characterized in that, The image enhancement processing of the conductive particle image to be detected in step S2 specifically involves: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected, eliminate noise and smooth the image; S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically: In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location; S23: Set the threshold constant for grayscale difference, if the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows: In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale value of pixels used to enhance the image; S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically: In the formula, A low threshold for the grayscale values ​​of pixels used to enhance the image; S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically: In the formula, This represents the grayscale value of a pixel after image enhancement.

3. The automatic detection method for the number and offset of conductive particles in a liquid crystal display screen according to claim 1, characterized in that, The specific steps for obtaining the bump region set in step S3 are as follows: S31: Project the conductive particle image along the Y-axis, calculate the sum of gray values ​​of each row of pixels in the conductive particle image, and obtain the set of sums of gray values ​​of pixels in the horizontal direction. S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories; S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

4. An automatic detection device for the number and misalignment of conductive particles in a liquid crystal display screen, characterized in that, The device includes an optical system and electronic equipment, the electronic equipment including a processor, a memory, and a computer program stored in the memory and executable on the processor. The device performs the following steps: S2; Perform image enhancement processing on the image of the conductive particles to be detected to improve the image contrast and highlight the conductive particles; S3: Segment the image of conductive particles after image enhancement to obtain a set of bump regions; S4: Calculate the number of conductive particles in the obtained bump region set; S5: Map the calculated conductive particles to the corresponding blank bump projection area and calculate the displacement of the conductive particles; S6: Output the number of conductive particles and the result of the offset; The specific steps in step S4 for calculating the number of conductive particles in the obtained bump region set are as follows: S41: Apply Gaussian filtering to the acquired bump region set to remove noise interference and smooth the image; S42: Binarize the set of bump regions processed by Gaussian filtering to obtain a binary image; S43: Perform an opening operation on the binary image by first erosion and then dilation. The convolution kernels for both the erosion and dilation operations are set to 3. S44: Locate the particle boundaries of all bump regions in the binary image, calculate the area of ​​the particle boundary for each bump region, and if the boundary area is less than 5 or greater than 200, determine that the particle is an interference particle and do not count it in the number of conductive particles in its respective bump region; for particles whose boundary area is... The particle boundary of the condition is defined, and the particle is counted as a conductive particle in the corresponding bump region; the number of conductive particles in each bump region is then calculated. The calculation of the displacement of the conductive particles in step S5 is specifically as follows: S51: Project all conductive particles within each bump region onto the selected corresponding bump region rectangle, and calculate the centroid of the region composed of all conductive particle projections. S52: Calculate the distance difference between the centroid and the center of the corresponding bump region rectangle, as follows: In the formula, the The distance difference between the centroid and the center of the corresponding bump region rectangle is the value of the distance difference between the centroid and the center of the corresponding bump region rectangle. For the centroid coordinates, the These are the coordinates of the center of the rectangle corresponding to the bump area; S53: Set a distance difference threshold. If the distance difference is less than the given threshold, it is determined that the conductive particles in the bump area are not misaligned. If the distance difference is greater than the given threshold, it is determined that the conductive particles in the bump area are misaligned, and the misalignment value is the distance difference.

5. The automatic detection device for the number and misalignment of conductive particles in a liquid crystal display screen according to claim 4, characterized in that, The image enhancement processing of the conductive particle image to be detected in step S2 specifically involves: S21: Use the set convolution kernel to perform Gaussian filtering on the image of the conductive particles to be detected, eliminate noise and smooth the image; S22: Traverse the pixels of the image of the conductive particles to be detected, and calculate the grayscale difference before and after Gaussian filtering for each pixel, specifically: In the formula, These are the pixel coordinates. The pixel coordinates in the images of the conductive particles to be detected before and after Gaussian filtering are shown. The grayscale difference at that location The pixel coordinates in the image of the conductive particles to be detected before Gaussian filtering. grayscale value at that location The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The grayscale value at that location; S23: Set the threshold constant for grayscale difference, if the grayscale difference... If the value is greater than the threshold, calculate the high threshold for the grayscale value of the pixel to be enhanced in the image, specifically as follows: In the formula, The pixel coordinates in the image of the conductive particles to be detected after Gaussian filtering. The transition compensation grayscale value at the location, Grayscale difference The threshold constant, A high threshold for the grayscale value of pixels used to enhance the image; S24: Calculate the low threshold of the grayscale values ​​of the pixels to be enhanced in the image, specifically: In the formula, A low threshold for the grayscale values ​​of pixels used to enhance the image; S25: Based on the high threshold With low threshold grayscale values ​​of pixels before Gaussian filtering Further processing yields the enhanced pixel grayscale values, specifically: In the formula, This refers to the grayscale value of the pixel after image enhancement.

6. The automatic detection device for the number and misalignment of conductive particles in a liquid crystal display screen according to claim 4, characterized in that, The specific steps for obtaining the bump region set in step S3 are as follows: S31: Project the conductive particle image along the Y-axis, calculate the sum of gray values ​​of each row of pixels in the conductive particle image, and obtain the set of sums of gray values ​​of pixels in the horizontal direction. S32: Perform K-means clustering on the set of sums of gray values ​​of the statistically analyzed horizontal pixels, specifying the number of clusters as 2, to divide the horizontal pixel region of the conductive particle image into 2 categories; S33: Based on the rule that the gray value of the bump region is lower than that of the non-bump region, the horizontal pixel regions of the two clustered categories are divided into bump regions and non-bump regions.

Images