A polishing pad groove detection method and system

Abstract

The present invention relates to the field of image analysis technology, and in particular to a method and system for detecting grooves of a polishing pad. The method comprises: obtaining a connected domain of a grayscale image of a polishing pad, performing skeleton line segmentation, and screening a target groove area according to a minimum circumscribed rectangle and area characteristics; dividing the target groove area into a groove defect area and a normal area according to the area difference with the normal area; determining a grayscale difference index according to the grayscale difference with the normal area; performing circle fitting on each groove defect area to obtain a fitting circle, and determining a discrete index according to the morphology of each groove defect area and the fitting circle; combining the grayscale difference index and the discrete index to determine the possibility of small holes, and screening small hole areas and abnormal groove areas according to the possibility of small holes. The present invention can accurately identify small hole areas and abnormal grooves, and improve the accuracy and reliability of polishing pad groove detection.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and in particular to a method and system for detecting grooves of a polishing pad. Background Art

[0002] Polishing pads are also called polishing leather, polishing cloth, and polishing sheet. They are important auxiliary materials that determine the surface quality in chemical mechanical polishing. According to whether they contain abrasives, they can be divided into polishing pads with abrasives and polishing pads without abrasives; according to different materials, they can be divided into polyurethane polishing pads, non-woven polishing pads, and composite polishing pads; according to different surface structures, they can be roughly divided into flat type, grid type, and spiral type.

[0003] During the polishing pad anomaly detection process, it is necessary to perform image analysis on the polishing pad grooves to realize image anomaly detection. The anomalies of the polishing pad grooves are mainly irregular textures and abnormal textures generated during the groove pressing process. Since there are many small holes distributed on the surface of the polishing pad, these small holes are normal material characteristics and have little effect on the quality of the polishing pad itself. The small holes will be identified as irregular abnormal grooves that actually affect the quality, thereby interfering with the image anomaly detection of the grooves and reducing the accuracy and reliability of the polishing pad groove image detection. Summary of the invention

[0004] In order to solve the technical problem that small holes in the related art are identified as irregular abnormal grooves that actually affect the quality, thereby interfering with the image abnormality detection of the grooves and reducing the accuracy and reliability of the polishing pad groove image detection, the present invention provides a polishing pad groove detection method and system, and the technical solutions adopted are as follows:

[0005] The present invention provides a method for detecting a polishing pad groove, the method comprising:

[0006] Acquire at least one connected domain in the polishing pad grayscale image of the grid-type polishing pad, segment according to the skeleton line of each connected domain, and determine the suspected polishing pad groove area;

[0007] Determine the groove possibility of the suspected polishing pad groove region according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove region and the area of ​​the suspected polishing pad groove region; and take the suspected polishing pad groove region whose groove possibility is greater than a preset first threshold as a target groove region;

[0008] Determine the defect possibility of any target groove region according to the area difference between any target groove region and all other target groove regions in the grayscale image of the polishing pad, and divide the target groove region into a groove defect region and a normal region according to the defect possibility; determine the grayscale difference index of the groove defect region according to the grayscale difference between any groove defect region and the normal region;

[0009] Perform circle fitting on all edge pixel points of each groove defect area to obtain a fitting circle for each suspected defect area, and determine the discrete index of the groove defect area and the fitting circle according to the distance between the edge pixel point of each groove defect area and the nearest point on the edge of the corresponding fitting circle;

[0010] The pinhole possibility of the groove defect area is determined by combining the grayscale difference index and the discrete index, and the groove defect area with the pinhole possibility greater than the preset second threshold is regarded as a pinhole area, otherwise it is regarded as an irregular abnormal groove area.

[0011] Furthermore, the segmentation according to the skeleton line of each connected domain to determine the suspected polishing pad groove area includes:

[0012] Use Zhang-Suen algorithm to operate each connected domain and obtain the skeleton line of each connected domain;

[0013] When there are more than or equal to 3 other pixels on the skeleton line within the eight neighborhoods of a pixel point on the skeleton line, the pixel point is used as a branch point;

[0014] The skeleton line is segmented at each branch point to obtain the suspected polishing pad groove area.

[0015] Further, the determining the groove possibility of the suspected polishing pad groove region according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove region and the area of ​​the suspected polishing pad groove region includes:

[0016] Determine the minimum bounding rectangle of each suspected polishing pad groove region based on a minimum bounding rectangle algorithm;

[0017] The aspect ratio of the minimum circumscribed rectangle is used as a first possible influence coefficient;

[0018] The ratio of the area of ​​the suspected polishing pad groove region to the area of ​​the minimum circumscribed rectangle is used as a second possible influence coefficient, wherein the area of ​​the minimum circumscribed rectangle is the product of the length and width of the minimum circumscribed rectangle;

[0019] The product of the first possibility influence coefficient and the second possibility influence coefficient is calculated, and normalized to obtain the groove possibility.

[0020] Furthermore, the determining the defect possibility of any target groove region according to the area difference between any target groove region and all other target groove regions in the polishing pad grayscale image includes:

[0021] The defect probability of any target groove area is obtained by normalizing the average of the absolute values ​​of the area differences between any target groove area and all other target groove areas.

[0022] Further, dividing the target groove area into a groove defect area and a normal area according to the defect possibility includes:

[0023] The target groove region whose defect probability is greater than a preset third threshold is regarded as a groove defect region;

[0024] The target groove region whose defect probability is less than or equal to a preset third threshold is regarded as a normal region.

[0025] Furthermore, the step of determining the grayscale difference index of the groove defect region according to the grayscale difference between any groove defect region and a normal region includes:

[0026] Calculate the average grayscale value of all pixels in the normal area to obtain a normal average value;

[0027] The absolute value of the difference between any pixel point in any groove defect area and the normal mean is used as the normal difference index of the pixel point;

[0028] The average of the normal difference indicators of all pixels in the same groove defect area is used as the grayscale difference indicator of the groove defect area.

[0029] Furthermore, performing circle fitting on all edge pixels of each groove defect area to obtain a fitting circle for each suspected defect area includes:

[0030] The least squares fitting circle is calculated for all edge pixels of each groove defect area to obtain a fitting circle for each suspected defect area.

[0031] Furthermore, the discrete index of the groove defect area and the fitting circle is determined according to the distance between the edge pixel point of each groove defect area and the nearest point of the corresponding fitting circle edge, including:

[0032] Determine the point with the smallest Euclidean distance between the edge pixel point of each groove defect area and the edge of the fitting circle as the closest point;

[0033] The Euclidean distance between the edge pixel and its nearest point is used as the fitting circle distance of the edge pixel;

[0034] The mean value of the fitting circle distance of all edge pixels in the groove defect area is taken as the discrete index between the groove defect area and the fitting circle.

[0035] Furthermore, the combining of the grayscale difference index and the discrete index to determine the possibility of pinholes in the groove defect area includes:

[0036] The opposite number of the discrete index is normalized to the maximum and minimum values ​​to obtain the small hole recognition index;

[0037] The product of the grayscale difference index and the pinhole identification index is calculated and normalized to be the pinhole possibility of the groove defect area.

[0038] On the other hand, a polishing pad groove detection system is also provided, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of a polishing pad groove detection method as described in any one of the preceding items when executing the computer program.

[0039] The present invention has the following beneficial effects:

[0040] The present invention determines the suspected polishing pad groove area by segmenting at least one connected domain in the polishing pad grayscale image of the grid-type polishing pad according to the skeleton line of each connected domain, and then performs groove analysis on the suspected polishing pad groove area according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove area and the area of ​​the suspected polishing pad groove area, so as to accurately obtain the target groove area belonging to the groove. However, since the small hole area will be mistakenly identified as the groove area and has a large difference from the normal groove area, it affects the abnormal identification of the groove area, resulting in the detection of the polishing pad groove being affected by the small hole area. Therefore, it is necessary to first analyze the groove defect area and combine it with the surface The groove defect area is obtained by integrating the features, and the grayscale difference is combined to determine the grayscale difference index of the groove defect area. The grayscale difference index is convenient for the subsequent grayscale distinction between the small hole area and the irregular abnormal groove area; then, a morphological analysis of circle fitting is performed on each groove defect area to determine the discrete index of the groove defect area and the fitting circle. The acquisition of the discrete index is convenient for the subsequent shape distinction between the small hole area and the irregular abnormal groove area; finally, the grayscale difference index and the discrete index are combined to screen the small hole area and the irregular abnormal groove area, so as to eliminate the influence of the small hole area on the image analysis and recognition of the abnormal groove, and improve the accuracy and reliability of the polishing pad groove image detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A flow chart of a method for detecting grooves of a polishing pad provided by one embodiment of the present invention;

[0043] Figure 2 A schematic diagram of a grayscale image of a polishing pad provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a polishing pad groove detection method and system according to the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0045] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0046] In order to enable those skilled in the art to understand the content of the present invention, some basic terms will be explained and elaborated in detail below.

[0047] Polishing pads are also called polishing leather, polishing cloth, and polishing sheet. They are important auxiliary materials that determine the surface quality in chemical mechanical polishing. According to whether they contain abrasives, they can be divided into polishing pads with abrasives and polishing pads without abrasives; according to different materials, they can be divided into polyurethane polishing pads, non-woven polishing pads, and composite polishing pads; according to different surface structures, they can be roughly divided into flat type, grid type, and spiral type. The polishing pad in the embodiment of the present invention is mainly a grid type polishing pad, that is, the grooves of the polishing pad are in a grid type. See Figure 2 , Figure 2 A schematic diagram of a grayscale image of a polishing pad provided by an embodiment of the present invention.

[0048] The specific scheme of the polishing pad groove detection method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0049] See also Figure 1 , which shows a flow chart of a polishing pad groove detection method provided by an embodiment of the present invention, the method comprising:

[0050] S101: acquiring at least one connected domain in a polishing pad grayscale image of a grid-type polishing pad, segmenting the connected domain according to a skeleton line of each connected domain, and determining a suspected polishing pad groove region.

[0051] In the embodiment of the present invention, a high-definition camera is installed on the production line to photograph the polishing pad to obtain a plurality of original images of the polishing pad, and each original image of the polishing pad is subjected to image preprocessing. The image preprocessing in the embodiment of the present invention can specifically include image denoising, image grayscale processing, semantic segmentation and other preprocessing methods.

[0052] Among them, the image denoising process can be specifically, for example, mean denoising process, and the image grayscale process can be specifically, for example, mean grayscale process. Of course, in other embodiments of the present invention, the technical means used for preprocessing can also be adjusted according to needs to meet the actual usage scenario, and no further elaboration or limitation is made on this. After the image preprocessing, a polishing pad grayscale image is obtained.

[0053] It should be noted that the polishing pad is usually made of polyurethane. Due to the characteristics of polyurethane itself, its surface contains holes formed by bubbles. The holes are similar to the abnormal characteristics of the polishing pad grooves, but in fact, the small holes are a normal phenomenon of polyurethane itself and have no effect on the quality of the polishing pad itself. However, since the small holes are identified as defective areas, they affect the image abnormality detection results of the polishing pad grooves. Therefore, it is necessary to conduct an independent analysis of the holes to determine the quality of the polishing pad grooves.

[0054] The grayscale values ​​of pixels in the polishing pad groove area and hole area are significantly smaller than those in other areas. A threshold segmentation algorithm can be used to operate on each polishing pad grayscale image to obtain several connected domains. Threshold segmentation is a technology well known to technicians in this field and is not limited to this.

[0055] After determining the connected domain, it is necessary to analyze the connected domain. Since the polishing pad grooves are in a grid shape, they can be divided into different small units for ease of analysis. Each small unit represents an edge in the grid, so that specific analysis can be performed on all the groove edges of the grid.

[0056] Furthermore, in some embodiments of the present invention, segmentation is performed based on the skeleton line of each connected domain to determine the suspected polishing pad groove area, including: using the Zhang-Suen algorithm to operate on each connected domain to obtain the skeleton line of each connected domain; when there are greater than or equal to 3 other pixel points on the skeleton line within the eight-neighborhood of a pixel point on the skeleton line, the pixel point is used as a branch point; and the skeleton line is segmented at each branch point to obtain the suspected polishing pad groove area.

[0057] The Zhang-Suen algorithm is used to operate each connected domain of each polishing pad grayscale image to obtain the skeleton line of each connected domain of each polishing pad grayscale image.

[0058] Each skeleton line is segmented at each branch point, and each connected domain is segmented into multiple regions, which are recorded as suspected polishing pad groove regions. The skeleton line at the branch point is defined as follows: when there are more than or equal to 3 pixels in the eight neighborhoods of a pixel point on the skeleton line, the pixel point is a branch point. The Zhang-Suen algorithm is a well-known technology, and the specific method is not introduced here.

[0059] Among them, there are more than or equal to 3 pixels in the eight neighborhoods, which means that there are at least three extended branches in the eight neighborhoods, further indicating that it is a mesh node position, which is taken as a branch point. Therefore, the two adjacent branch points are segmented, and the connected domain area between the two branch points is taken as the suspected polishing pad groove area. It can be understood that if the connected domain between the two branch points belongs to a connected domain, the entire connected domain is directly cut off from the branch point position to obtain a suspected polishing pad groove area, and when the two adjacent branch points belong to different connected domains, each branch point is cut off separately to obtain different suspected polishing pad groove areas.

[0060] S102: Determine the groove possibility of the suspected polishing pad groove area according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove area and the area of ​​the suspected polishing pad groove area; and take the suspected polishing pad groove area whose groove possibility is greater than a preset first threshold as the target groove area.

[0061] Among them, since the polishing pad grooves are in a grid shape, the grooves between two normal branch points should be an area that is approximately a long rectangular strip. Therefore, the minimum circumscribed rectangle can be used for specific analysis to determine the possibility of grooves in the suspected polishing pad groove area.

[0062] Further, in some embodiments of the present invention, the groove possibility of the suspected polishing pad groove area is determined based on the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove area and the area of ​​the suspected polishing pad groove area, including: determining the minimum circumscribed rectangle of each suspected polishing pad groove area based on the minimum circumscribed rectangle algorithm; using the length-to-width ratio of the minimum circumscribed rectangle as the first possibility influence coefficient; using the ratio of the area of ​​the suspected polishing pad groove area to the area of ​​the minimum circumscribed rectangle as the second possibility influence coefficient, wherein the area of ​​the minimum circumscribed rectangle is the product of the length and width of the minimum circumscribed rectangle; calculating the product of the first possibility influence coefficient and the second possibility influence coefficient, and normalizing them to obtain the groove possibility.

[0063] Among them, since the groove between two normal branch points is an approximately long rectangular area, its length-to-width ratio is relatively large, that is, the larger the first possibility influence coefficient is, the more it conforms to the normal groove characteristics.

[0064] At the same time, since the normal groove edge is straight, when constructing the minimum circumscribed rectangle, it is usually close to the straight groove edge, that is, the ratio of the area of ​​the suspected polishing pad groove area to the area of ​​the minimum circumscribed rectangle is large. The larger the value of the second possibility influence coefficient, the more it conforms to the normal groove characteristics.

[0065] Therefore, in the embodiment of the present invention, the product of the first possibility influence coefficient and the second possibility influence coefficient is calculated, and the groove possibility is obtained by normalization. The corresponding calculation formula is:

[0066]

[0067] in, Indicates Grayscale image of the polishing pad groove likelihood in the suspected polishing pad groove area; represents the linear normalization function; Indicates Grayscale image of the polishing pad The length of the longest side of the smallest circumscribed rectangle of the suspected polishing pad groove area; Indicates Grayscale image of the polishing pad The length of the shortest side of the smallest circumscribed rectangle of the suspected polishing pad groove area; Indicates Grayscale image of the polishing pad The area of ​​the suspected polishing pad groove area is .

[0068] It should be noted that Grayscale image of the polishing pad The larger the ratio of the longest side to the shortest side of the minimum circumscribed rectangle of the suspected polishing pad groove area, the Grayscale image of the polishing pad The more the suspected polishing pad groove area tends to be elongated, the Grayscale image of the polishing pad The greater the possibility that the first suspected polishing pad groove area is the target groove area; Grayscale image of the polishing pad The larger the area of ​​the suspected polishing pad groove area, the Grayscale image of the polishing pad The greater the probability that the suspected polishing pad groove area is the target groove area.

[0069] The preset first threshold is the threshold value of the groove possibility. Specifically, the preset first threshold may be 0.8, that is, the suspected polishing pad groove area with a groove possibility greater than 0.8 is taken as the target groove area. The suspected polishing pad groove area with a groove possibility less than or equal to 0.8 may be a relatively obvious small hole area or a relatively obvious abnormal area, which is very easy to be identified in the subsequent abnormality detection process. Therefore, this solution does not specifically analyze it. This solution only analyzes the abnormal state disguised as the target groove area, so that the overall image analysis is more accurate.

[0070] Among them, the target groove area, that is, after area analysis, is determined to be a preliminarily normal polishing pad groove area, but this type of "normal polishing pad groove area" still contains successfully disguised holes and irregular long strip textures, so further defect analysis of the target groove area is required.

[0071] S103: Determine the defect possibility of any target groove area according to the area difference between any target groove area and all other target groove areas in the grayscale image of the polishing pad, and divide the target groove area into a groove defect area and a normal area according to the defect possibility; determine the grayscale difference index of the groove defect area according to the grayscale difference between any groove defect area and the normal area.

[0072] Among them, the groove defect area includes the area where small holes are formed and the area caused by actual defects. The area caused by actual defects is usually irregular groove texture and damage caused by the groove during the printing process. The small hole area is a normal phenomenon on the polishing pad, and it has little effect on the quality of the polishing pad; while the irregular groove texture and damage are abnormal phenomena and will affect the quality of the overall polishing pad. Therefore, it is necessary to screen the groove defect area and conduct further feature analysis on the groove defect area.

[0073] First, the groove defect area is screened:

[0074] Among them, since the areas of a normal trench region and a trench region with defects are greatly different, the possibility of defects is first analyzed based on the area difference.

[0075] Furthermore, in some embodiments of the present invention, the defect possibility of any target groove area is determined based on the area difference between any target groove area and all other target groove areas in the grayscale image of the polishing pad, including: normalizing the average of the absolute values ​​of the area differences between any target groove area and all other target groove areas to obtain the defect possibility of any target groove area.

[0076] Among them, the absolute value of the area difference between any target groove area and each other target groove area is selected as the area difference value corresponding to the two target groove areas, and the mean of all area difference values ​​corresponding to any target groove area is linearly normalized to obtain the defect possibility of the target groove area. That is, the larger the area difference between the target groove area and all other target groove areas, the more abnormal the area of ​​the target groove area is, and the greater the defect possibility is.

[0077] All target groove areas can be classified according to defect possibilities, and groove defect areas with abnormal performance can be screened out.

[0078] Furthermore, in some embodiments of the present invention, the target groove area is divided into a groove defect area and a normal area according to the defect possibility, including: treating the target groove area with a defect possibility greater than a preset third threshold as a groove defect area; treating the target groove area with a defect possibility less than or equal to the preset third threshold as a normal area.

[0079] Among them, the preset third threshold is the threshold value of the defect possibility. Optionally, the preset third threshold can be 0.5, that is, the target groove area with a defect possibility greater than 0.5 is regarded as a groove defect area; the target groove area with a defect possibility less than or equal to 0.5 is regarded as a normal area.

[0080] Secondly, further characterization analysis is performed on the trench defect area:

[0081] It should be noted that the small holes are only distributed on the surface, because the irregular abnormal groove area is a deep texture caused by machine die-casting errors, while the small holes are usually shallow textures caused by bubbles. The depth of the small holes is shallower than that of the irregular abnormal groove area. The grayscale difference between the irregular abnormal groove area and the normal groove area is small. Therefore, further analysis can be performed based on the grayscale.

[0082] Furthermore, in some embodiments of the present invention, a grayscale difference index of a groove defect area is determined based on the grayscale difference between any groove defect area and a normal area, including: calculating the mean grayscale value of pixels in all normal areas to obtain a normal mean; using the absolute value of the difference between any pixel in any groove defect area and the normal mean as the normal difference index of the pixel; and using the mean of the normal difference indicators of all pixels in the same groove defect area as the grayscale difference index of the groove defect area.

[0083] In this solution, the grayscale value average of all pixels in the normal area is used to identify the small hole area and the irregular abnormal groove area.

[0084] In an embodiment of the present invention, the absolute value of the difference between any pixel point in any groove defect area and the normal mean is calculated as the normal difference index of the pixel point; the larger the normal difference index is, the more it conforms to the shallow grayscale characteristics of the small hole area, and the smaller the difference index is, the more it conforms to the grayscale characteristics of the irregular abnormal groove area.

[0085] Therefore, the average of the normal difference indexes of all pixels in the same groove defect area is used as the grayscale difference index of the groove defect area. The grayscale difference index represents the overall characteristics of a groove defect area. By integrating all pixels in the same groove defect area, the grayscale difference index of the groove defect area is obtained. The larger the grayscale difference index, the more it conforms to the characteristics of the pinhole area, and the smaller the grayscale difference index, the more it conforms to the characteristics of the irregular abnormal groove area with texture printing errors.

[0086] S104: performing circle fitting on all edge pixel points of each groove defect area to obtain a fitting circle for each suspected defect area, and determining a discrete index between the groove defect area and the fitting circle according to the distance between the edge pixel point of each groove defect area and the nearest point on the edge of the corresponding fitting circle.

[0087] Among them, since the small hole area is a shallow texture caused by the influence of bubbles, it presents hole characteristics as a whole, that is, it is more consistent with the circular feature. Based on this, the small hole area is further identified.

[0088] Furthermore, in some embodiments of the present invention, a circle is fitted on all edge pixel points of each groove defect area to obtain a fitting circle for each suspected defect area, including: using a least squares fitting circle to operate on all edge pixel points of each groove defect area to obtain a fitting circle for each suspected defect area.

[0089] The least square method is used to obtain the fitting circle, which is well known to those skilled in the art and is not limited thereto.

[0090] In an embodiment of the present invention, since the small hole area is closer to the fitting circle, it is analyzed. Further, in some embodiments of the present invention, the discrete index of the groove defect area and the fitting circle is determined according to the distance between the edge pixel point of each groove defect area and the nearest point on the edge of the corresponding fitting circle, including: determining the point with the smallest Euclidean distance value between the edge pixel point of each groove defect area and the edge of the fitting circle as the nearest point; taking the Euclidean distance between the edge pixel point and its nearest point as the fitting circle distance of the edge pixel point; taking the average of the fitting circle distances of all edge pixel points of the groove defect area as the discrete index of the groove defect area and the fitting circle.

[0091] Among them, the small hole area is closer to the fitting circle, that is, each edge pixel point in the small hole area is closer to the point on the fitting circle. Therefore, the mean of the fitting circle distances of all edge pixels in the groove defect area is used as the discrete index of the groove defect area and the fitting circle. The smaller the discrete index, the more it conforms to the shape characteristics of the small hole area, and the larger the discrete index, the more it conforms to the shape characteristics of the irregular abnormal groove area.

[0092] S105: Determine the pinhole possibility of the groove defect area by combining the grayscale difference index and the discrete index, and regard the groove defect area with the pinhole possibility greater than the preset second threshold as the pinhole area, otherwise as the irregular abnormal groove area.

[0093] Since both the degree difference index and the discrete index can distinguish the small hole characteristics from the irregular abnormal groove characteristics, the small hole defect area and the irregular abnormal groove area can be identified from the groove defect area.

[0094] Furthermore, in some embodiments of the present invention, the possibility of pinholes in the groove defect area is determined by combining the grayscale difference index and the discrete index, including: normalizing the opposite of the discrete index to the maximum and minimum values ​​to obtain a pinhole recognition index; calculating the product of the grayscale difference index and the pinhole recognition index, and normalizing it as the possibility of pinholes in the groove defect area.

[0095] Among them, since the smaller the discrete index is, the more it conforms to the shape characteristics of the small hole area, and the larger the discrete index is, the more it conforms to the shape characteristics of the irregular abnormal groove area, it is necessary to perform negative correlation processing on the discrete index, and normalize the opposite number of the discrete index to the maximum and minimum values ​​to obtain the small hole recognition index. The larger the small hole recognition index is, the more it conforms to the shape characteristics of the small hole area.

[0096] Since the larger the grayscale difference index is, the more it conforms to the grayscale characteristics of the pinhole area, in the embodiment of the present invention, the product of the grayscale difference index and the pinhole identification index is directly calculated and normalized as the pinhole possibility of the groove defect area.

[0097] Among them, the preset second threshold is the threshold value of the possibility of a small hole. In the embodiment of the present invention, the preset second threshold can be specifically 0.8, that is, the groove defect area with a small hole possibility greater than 0.8 is regarded as a small hole area, and the groove defect area with a small hole possibility less than or equal to 0.8 is regarded as an irregular abnormal groove area.

[0098] Thus, the influence of the pinhole area is eliminated and the abnormal groove area with irregular shape is accurately determined.

[0099] The present invention determines the suspected polishing pad groove area by segmenting at least one connected domain in the polishing pad grayscale image of the grid-type polishing pad according to the skeleton line of each connected domain, and then performs groove analysis on the suspected polishing pad groove area according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove area and the area of ​​the suspected polishing pad groove area, so as to accurately obtain the target groove area belonging to the groove. However, since the small hole area will be mistakenly identified as the groove area and has a large difference from the normal groove area, it affects the abnormal identification of the groove area, resulting in the detection of the polishing pad groove being affected by the small hole area. Therefore, it is necessary to first analyze the groove defect area. The groove defect area is obtained by combining the area characteristics, and the grayscale difference is combined to determine the grayscale difference index of the groove defect area. The grayscale difference index is convenient for the subsequent grayscale distinction between the small hole area and the irregular abnormal groove area; then, a morphological analysis of circle fitting is performed on each groove defect area to determine the discrete index of the groove defect area and the fitting circle. The acquisition of the discrete index is convenient for the subsequent shape distinction between the small hole area and the irregular abnormal groove area; finally, the grayscale difference index and the discrete index are combined to screen the small hole area and the irregular abnormal groove area, so as to eliminate the influence of the small hole area on the recognition of the abnormal groove, and improve the accuracy and reliability of the polishing pad groove detection.

[0100] The present invention also provides a polishing pad groove detection system, the system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the polishing pad groove detection method as described above are implemented.

[0101] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0102] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting grooves of a polishing pad, characterized in that: The method comprises: Acquire at least one connected domain in the polishing pad grayscale image of the grid-type polishing pad, segment according to the skeleton line of each connected domain, and determine the suspected polishing pad groove area; Determine the groove possibility of the suspected polishing pad groove area according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove area and the area of ​​the suspected polishing pad groove area; take the suspected polishing pad groove area with the groove possibility greater than a preset first threshold as the target groove area, wherein the target groove area is a polishing pad groove area determined to be preliminarily normal after area analysis, and contains holes and irregular long strip textures; Determine the defect possibility of any target groove region according to the area difference between any target groove region and all other target groove regions in the grayscale image of the polishing pad, and divide the target groove region into a groove defect region and a normal region according to the defect possibility; determine the grayscale difference index of the groove defect region according to the grayscale difference between any groove defect region and the normal region; Perform circle fitting on all edge pixel points of each groove defect area to obtain a fitting circle for each suspected defect area, and determine the discrete index of the groove defect area and the fitting circle according to the distance between the edge pixel point of each groove defect area and the nearest point on the edge of the corresponding fitting circle; Determine the pinhole possibility of the groove defect area by combining the grayscale difference index and the discrete index, and regard the groove defect area with the pinhole possibility greater than the preset second threshold as the pinhole area, otherwise as the irregular abnormal groove area; The step of combining the grayscale difference index and the discrete index to determine the possibility of pinholes in the groove defect area includes: The opposite number of the discrete index is normalized to the maximum and minimum values ​​to obtain the small hole recognition index; The product of the grayscale difference index and the pinhole identification index is calculated and normalized to be the pinhole possibility of the groove defect area.

2. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: The segmentation according to the skeleton line of each connected domain to determine the suspected polishing pad groove area includes: Use Zhang-Suen algorithm to operate each connected domain and obtain the skeleton line of each connected domain; When there are more than or equal to 3 other pixels on the skeleton line within the eight neighborhoods of a pixel point on the skeleton line, the pixel point is used as a branch point; The skeleton line is segmented at each branch point to obtain the suspected polishing pad groove area.

3. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: The determining the groove possibility of the suspected polishing pad groove region according to the length and width of the minimum circumscribed rectangle of each suspected polishing pad groove region and the area of ​​the suspected polishing pad groove region includes: Determine the minimum bounding rectangle of each suspected polishing pad groove region based on a minimum bounding rectangle algorithm; The aspect ratio of the minimum circumscribed rectangle is used as a first possible influence coefficient; The ratio of the area of ​​the suspected polishing pad groove region to the area of ​​the minimum circumscribed rectangle is used as a second possible influence coefficient, wherein the area of ​​the minimum circumscribed rectangle is the product of the length and width of the minimum circumscribed rectangle; The product of the first possibility influence coefficient and the second possibility influence coefficient is calculated, and normalized to obtain the groove possibility.

4. A method for detecting a polishing pad groove according to claim 1, characterized in that: Determining the defect possibility of any target groove area according to the area difference between any target groove area and all other target groove areas in the polishing pad grayscale image includes: The defect probability of any target groove area is obtained by normalizing the average of the absolute values ​​of the area differences between any target groove area and all other target groove areas.

5. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: The step of dividing the target groove region into a groove defect region and a normal region according to the defect possibility comprises: The target groove region whose defect probability is greater than a preset third threshold is regarded as a groove defect region; The target groove region whose defect probability is less than or equal to a preset third threshold is regarded as a normal region.

6. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: Determining the grayscale difference index of the groove defect area according to the grayscale difference between any groove defect area and the normal area includes: Calculate the average grayscale value of all pixels in the normal area to obtain a normal average value; The absolute value of the difference between any pixel point in any groove defect area and the normal mean is used as the normal difference index of the pixel point; The average of the normal difference indicators of all pixels in the same groove defect area is used as the grayscale difference indicator of the groove defect area.

7. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: The step of performing circle fitting on all edge pixel points of each groove defect region to obtain a fitting circle for each suspected defect region includes: The least squares fitting circle is calculated for all edge pixels of each groove defect area to obtain a fitting circle for each suspected defect area.

8. A method for detecting grooves of a polishing pad as claimed in claim 1, characterized in that: The method of determining the discrete index of the groove defect area and the fitting circle according to the distance between the edge pixel point of each groove defect area and the nearest point of the corresponding fitting circle edge includes: Determine the point with the smallest Euclidean distance between the edge pixel point of each groove defect area and the edge of the fitting circle as the closest point; The Euclidean distance between the edge pixel and its nearest point is used as the fitting circle distance of the edge pixel; The mean value of the fitting circle distance of all edge pixels in the groove defect area is taken as the discrete index between the groove defect area and the fitting circle.

9. A polishing pad groove detection system, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images