Wafer group integrity detection method based on multi-line laser

Abstract

The invention discloses a wafer group integrity detection method based on multi-line laser, and relates to the field of wafer integrity detection, and the wafer group comprises a plurality of wafers which are vertically arranged at intervals along a first direction; when the integrity of the wafer group is detected, a plurality of laser devices are used for generating multiple beams of line laser which are arranged at intervals in the second direction and are parallel to one another, and each beam of line laser irradiates the edges of all wafers in the wafer group; collecting a detection image, wherein the detection image comprises all light spots formed by reflecting the multi-beam laser by the edges of the plurality of wafers; preprocessing the detection image to obtain a preprocessed image, acquiring light spot features, and screening according to the light spot features to obtain effective light spots; and performing feature extraction on all the effective light spots to obtain feature point coordinates of each effective light spot in the image coordinate system, forming a wafer group distribution diagram based on the feature point coordinates of the effective light spots, and judging the integrity of the wafer group according to the wafer group distribution diagram. According to the method, the integrity of the wafer can be efficiently and accurately detected in a complex environment.

Description

Technical Field

[0001] The present invention relates to the technical field of wafer group integrity detection, and in particular to a wafer group integrity detection method based on multi-line laser. Background Art

[0002] During the semiconductor manufacturing process, wafers need to go through multiple process steps, such as silicon wafer preparation, oxidation, photolithography, etching, doping, metallization and packaging. These steps are repeated continuously to manufacture multi-level, multi-functional integrated circuits. In order to remove impurities and contaminants generated during the manufacturing process, the wafers are usually cleaned after photolithography, etching and chemical mechanical polishing to ensure the quality and accuracy of subsequent processes.

[0003] When cleaning wafers, they are usually placed in a wafer basket with intervals of 25 or 50 wafers per group. The wafer basket is used to carry and transport wafers during the cleaning process. During the cleaning process, some uncertain factors may cause the wafer to fall off the wafer basket, or the local edge of the wafer may be damaged, destroying the integrity of the wafer group in the wafer basket. At this time, the abnormal situation of the wafer must be discovered in time, otherwise it will affect the subsequent wafer cleaning.

[0004] Currently, the inspection method for the integrity of the wafer group in the wafer basket usually relies on contact measurement equipment or optical inspection with a single light source. The traditional wafer integrity inspection method has the following limitations:

[0005] (1) Traditional optical detection is easily affected by environmental reflections. Since reflective surfaces such as water and mirrors often exist in the environment, a large amount of non-point reflection interference will be generated, resulting in inaccurate detection results.

[0006] (2) Traditional inspection methods often require going through all wafers in a wafer group one by one, which results in low inspection efficiency and cannot meet the needs of large-scale inspection.

[0007] (3) Traditional detection methods require complex hardware structures and a stable working environment, cannot meet the application scenarios that require rapid deployment and reconfiguration, and have poor application flexibility.

[0008] In summary, traditional detection methods are not sufficiently applicable under changing production environments and requirements for efficient batch testing. Summary of the invention

[0009] In view of the above problems and technical requirements, the inventors have proposed a wafer group integrity detection method based on multi-line laser. The technical solution of the present invention is as follows:

[0010] A wafer group integrity detection method based on multi-line laser, wherein the wafer group includes a plurality of wafers arranged vertically and spaced apart along a first direction;

[0011] When detecting the integrity of the wafer group, a plurality of lasers are used to generate a plurality of line laser beams that are arranged at intervals and parallel to each other along a second direction, and each line laser beam irradiates the edges of all wafers in the wafer group, and the second direction is perpendicular to the first direction;

[0012] Collecting a detection image, wherein the detection image includes all light spots formed by the reflection of multiple laser beams from the edges of multiple wafers;

[0013] The detected image is preprocessed to obtain a preprocessed image, and after the preprocessing, the light spot in the preprocessed image is extracted, the light spot feature is obtained, and the effective light spot is screened according to the light spot feature;

[0014] Feature extraction is performed on all effective light spots to obtain the feature point coordinates of each effective light spot in the image coordinate system, a wafer group distribution map is formed based on the feature point coordinates of all effective light spots, and the integrity of the wafer group is judged according to the wafer group distribution map.

[0015] A further technical solution is that the image coordinate system is established with any pixel point of the preprocessed image as the origin, the second direction as the x-axis direction, and the first direction as the y-axis direction;

[0016] The characteristic points of all the effective light spots are plotted in an image coordinate system to form a wafer group distribution map, and the integrity of the wafer group is judged according to the wafer group distribution map, including:

[0017] Setting a plurality of y-axis coordinate threshold ranges, wherein the y-axis coordinate threshold ranges correspond one to one to the wafers in the wafer group;

[0018] The effective spot feature points are divided into a plurality of effective spot feature point groups according to a plurality of y-axis coordinate threshold ranges, and the number of effective spot feature points in each effective spot feature point group is compared with the integrity threshold to judge the integrity of the wafer group.

[0019] A further technical solution is that when acquiring the light spot characteristics, the geometric characteristics, brightness characteristics and position characteristics of the light spot are acquired;

[0020] When obtaining an effective light spot by screening the light spot characteristics, the light spot whose geometric characteristics, brightness characteristics and position characteristics meet the effective light spot characteristics is taken as the effective light spot;

[0021] When obtaining the geometric features of the light spot, it includes calculating the area of ​​the light spot, the circularity C, and the aspect ratio of the minimum circumscribed rectangle; the circularity C of the light spot=4πA / P2, where A is the area of ​​the light spot and P is the perimeter of the light spot.

[0022] A further technical solution is that when preprocessing the detection image, the detection image is grayed;

[0023] When obtaining the brightness characteristics of the light spot, it includes obtaining the average gray value, contrast and brightness uniformity of the light spot, among which,

[0024] The average grayscale value of the light spot is the average grayscale value of all pixels in the light spot, the contrast of the light spot is the difference between the average grayscale value of the light spot and the average grayscale value of the background in the preprocessed image, and the uniform brightness value of the light spot is the ratio of the maximum grayscale value to the minimum grayscale value of the pixels in the light spot.

[0025] A further technical solution is that when obtaining the position characteristics of the light spot, it includes:

[0026] Obtain the minimum vertical distance between the edge of the light spot and each edge of the preprocessed image;

[0027] Obtain the coordinates of the centroid of each light spot in the preprocessed image in the image coordinate system, and calculate the distance between the centroids of adjacent light spots;

[0028] Dividing the spot centroid coordinates into a plurality of spot centroid coordinate groups according to a plurality of y-axis coordinate threshold ranges, and performing straight line fitting on the spot centroid coordinates in each spot centroid coordinate group to obtain a fitting straight line;

[0029] The residual between the centroid coordinates of each spot in each spot centroid coordinate group and the fitting straight line is calculated.

[0030] A further technical solution is that when the geometric features meet the effective light spot features, the area of ​​the light spot is within the effective light spot area range, the circularity C of the light spot is not less than the circularity threshold, and the aspect ratio of the minimum circumscribed rectangle of the light spot is within the effective light spot aspect ratio range;

[0031] When the brightness characteristics meet the effective light spot characteristics, they include: the average gray value of the light spot is not less than the average gray value threshold, the contrast of the light spot is not less than the contrast threshold, and the brightness uniformity of the light spot is not greater than the brightness uniformity threshold.

[0032] A further technical solution is that when the position feature meets the effective light spot feature, it includes:

[0033] The spacing between adjacent spot centroids is within the effective spot spacing range, the minimum vertical distance between the spot edge and each edge of the preprocessed image is not less than the edge threshold, and the residual between the spot centroid coordinates and the corresponding fitting straight line is not greater than the residual threshold.

[0034] Its further technical solution is, when extracting features from any effective light spot, it includes determining the edge points of the effective light spot through an edge detection algorithm, marking the connected edge points as a connected domain through connected domain analysis, and using the centroid coordinates of the connected domain as the feature point coordinates of the effective light spot.

[0035] A further technical solution is that the preprocessing of the detection image also includes performing noise reduction processing on the grayscale detection image, and the noise reduction processing at least includes performing morphological filtering on the grayscale detection image.

[0036] A further technical solution is that the centers of multiple wafers in the wafer group overlap, and during detection, the multiple wafers in the wafer group remain still or rotate with the center of the wafer as the rotation center.

[0037] The beneficial technical effects of the present invention are:

[0038] The present invention proposes a non-contact wafer group integrity detection method, which uses multi-beam line lasers to irradiate the edge of the wafer, and judges the integrity of the wafer group by collecting the spot distribution formed by the reflected line lasers at the edge of the wafer. It has batch detection capabilities, high detection efficiency, and the required device settings are simple, and it is suitable for multiple application scenarios. The present invention pre-processes the collected detection images, and selects effective light spots for integrity analysis based on the characteristics of the light spots, which can effectively eliminate the reflection interference caused by the external environment, especially in an environment with a water surface or a mirror plane, it can still maintain efficient and accurate detection capabilities. The combination of multi-line laser irradiation and visual analysis not only improves the detection speed, but also greatly improves the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the layout of the laser and the camera in one embodiment of the present invention.

[0040] Figure 2 It is a flowchart of an embodiment of a wafer group integrity detection method based on multi-line laser provided by the present invention.

[0041] Figure 3 It is a side schematic diagram of multiple laser beams irradiating edges of multiple wafers in one embodiment of the present invention.

[0042] Figure 4 In an embodiment of the present invention, ideally Figure 3 Corresponding wafer group distribution map.

[0043] Figure 5 This is an embodiment of the wafer group distribution diagram under actual conditions provided by the present invention.

[0044] Reference numerals: 1-wafer assembly, 2-laser, 3-camera. DETAILED DESCRIPTION

[0045] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings.

[0046] The present invention proposes a wafer group integrity detection method based on multi-line laser, wherein the wafer group 1 comprises a plurality of wafers arranged vertically and spaced apart along a first direction;

[0047] When detecting the integrity of the wafer group 1, a plurality of lasers 2 are used to generate a plurality of line lasers that are arranged at intervals and parallel to each other along a second direction, and each line laser irradiates the edges of all wafers in the wafer group 1, wherein the second direction is perpendicular to the first direction;

[0048] Collecting a detection image, wherein the detection image includes all light spots formed by the reflection of multiple laser beams from the edges of multiple wafers;

[0049] The detected image is preprocessed to obtain a preprocessed image, and after the preprocessing, the light spot in the preprocessed image is extracted, the light spot feature is obtained, and the effective light spot is screened according to the light spot feature;

[0050] Feature extraction is performed on all effective light spots to obtain the feature point coordinates of each effective light spot in the image coordinate system, a wafer group distribution map is formed based on the feature point coordinates of all effective light spots, and the integrity of the wafer group is judged according to the wafer group distribution map.

[0051] As can be seen from the background technology, in semiconductor manufacturing, wafers are generally placed in a wafer basket in groups of 25 or 50. To ensure that the wafers do not contact or collide with each other during cleaning and other processes, the multiple wafers in the wafer basket are arranged at intervals. In general, the multiple wafers in the wafer group 1 are arranged at equal intervals, the centers of the wafers, i.e., the centers of the circles, coincide with each other, and the wafers are of the same size, ultimately forming a Figure 1 The cylindrical wafer assembly 1 shown in the figure, the first direction is Figure 1 The y-axis direction shown, the second direction is Figure 1 The x-axis direction is shown.

[0052] Specifically, the wafer group 1 placed in the wafer basket may fall off the wafer basket as a whole due to external factors during the semiconductor manufacturing process. Detecting the integrity of the wafer group 1 includes detecting whether a whole wafer has fallen off in the wafer group 1. In this embodiment, an 11-line laser 2 with a power of 50mW is used to emit 11 beams of laser light, which are irradiated to the edge of the wafer at intervals along the x-axis direction. Each beam of laser light irradiates the edge of all wafers, and multiple beams of laser light are evenly arranged on one side of the wafer group 1. A high-speed camera 3 is used to collect the detection image, and a wide-angle and distortion-free camera lens is selected to ensure the imaging quality of the detection image.

[0053] In this embodiment, the layout of the laser 2 and the camera 3 is as follows: Figure 1As shown, the laser 2 and the camera 3 are arranged on the same side of the wafer group 1. In this embodiment, the center of the camera 3 coincides with the center of the cylindrical side formed by the wafer group 1, and the camera needs to be able to capture clear line laser reflection spots on the reflection path of the line laser reflected by all wafers, even if the light spots reflected by the edges of the wafers in the wafer group 1 are all within the viewfinder of the camera 3. Specifically, the relative position of the laser 2 and the camera 3, the focal length of the camera lens, and the number and distribution density of the line lasers need to be determined according to the installation space provided in actual application. For example, when the relative distance that the camera 3 and the side of the wafer group 1 can be set is limited, resulting in a limited viewfinder range of the viewfinder, the distribution density of the laser can be increased, that is, the spacing between the line lasers can be reduced, or the number of line lasers can be reduced, so as to collect all the light spots reflected by the wafer. In specific implementation, the relative position of the laser 2 and the camera 3, the focal length of the camera lens, and the number and distribution density of the line lasers can be selected according to the actual situation, so as to meet the requirement that multiple line lasers are irradiated to the side of the wafer in parallel, and the camera 3 can collect all the light spots reflected by the wafer.

[0054] Figure 2 A flowchart of an embodiment of a wafer group integrity detection method is shown. Figure 2 As shown, after the detection image is collected, the detection image is preprocessed to obtain a preprocessed image. After preprocessing, the light spot in the preprocessed image is extracted, the light spot features are obtained, and the effective light spot is obtained according to the light spot features. That is, only the light spot that meets the characteristics of the laser reflection point is retained to eliminate the reflection interference caused by the external environment and improve the accuracy of the detection. After that, all effective light spots are feature extracted to obtain the feature point coordinates of each effective light spot in the image coordinate system, and a wafer group distribution map is formed based on the feature point coordinates of all effective light spots, and the integrity of the wafer group is judged according to the wafer group distribution map. The following describes the specific methods of image preprocessing, effective light spot screening, obtaining feature point coordinates, and judging the integrity of the wafer group.

[0055] Specifically, during the pre-processing, the detection image is first grayed, that is, the collected color image is converted into a gray image to reduce the complexity of the data. In this embodiment, the image is converted into an 8-bit gray image. Thereafter, the grayed detection image is subjected to noise reduction processing, and the noise reduction processing at least includes morphological filtering of the grayed detection image. Morphological filtering is suitable for eliminating noise and interference in the image. Through morphological filtering, non-point-shaped reflection spots with large areas and strong continuity in the detection image can be effectively removed to preliminarily eliminate external reflection interference. Morphological filtering includes opening and closing operations on the grayed detection image. Generally, the opening operation is performed first and then the closing operation is performed. The opening operation includes first corroding the detection image and then dilating it to remove small and isolated noise points in the detection image and smooth the boundaries of the light spot. The closing operation includes first dilating the detection image and then corroding it to fill the defects in the area of ​​the light spot, enhance the connectivity of the light spot area, and make the point light spot more obvious. The definition and specific implementation of the expansion treatment and corrosion treatment are consistent with the prior art and will not be elaborated here.

[0056] Preferably, the noise reduction process also includes smoothing the grayscale detection image using Gaussian filtering or median filtering before performing morphological filtering to reduce high-frequency noise in the image and further improve the accuracy of detection.

[0057] Furthermore, the spot area is extracted from the preprocessed image and the spot features are obtained. The common image segmentation method can be used to extract the spot area. The spot features are obtained by obtaining the geometric features, brightness features and position features of the spot. When the effective spot is obtained by screening the spot features, the spot whose geometric features, brightness features and position features all meet the effective spot features is taken as the effective spot, that is, the spot is screened from the three feature dimensions of geometric features, brightness features and position features.

[0058] Specifically, when obtaining the geometric features of the light spot, it includes calculating the area of ​​the light spot, the circularity C, and the aspect ratio of the minimum circumscribed rectangle; the circularity C of the light spot = 4πA / P2, where A is the area of ​​the light spot and P is the perimeter of the light spot. The geometric features meet the effective light spot features, that is, the area of ​​the light spot is within the effective light spot area range, the circularity C of the light spot is not less than the circularity threshold, and the aspect ratio of the minimum circumscribed rectangle of the light spot is within the effective light spot aspect ratio range. In this embodiment, the effective light spot area range is 30-100 square pixels, the circularity threshold is 0.85, and the effective light spot aspect ratio range is 0.8-1.2.

[0059] When obtaining the brightness characteristics of the light spot, it includes obtaining the average grayscale value, contrast and brightness uniformity value of the light spot, wherein the average grayscale value of the light spot is the average grayscale value of all pixels in the light spot, the contrast of the light spot is the difference between the average grayscale value of the light spot and the average grayscale value of the background in the preprocessed image, the background refers to the area other than the light spot area in the preprocessed image, and the brightness uniformity value of the light spot is the ratio of the maximum grayscale value to the minimum grayscale value of the pixel in the light spot. The brightness characteristics meet the characteristics of the effective light spot, that is, the average grayscale value of the light spot is not less than the average grayscale value threshold, the contrast of the light spot is not less than the contrast threshold, and the brightness uniformity value of the light spot is not greater than the brightness uniformity value threshold. In this embodiment, the average grayscale value threshold is 180, the contrast threshold is 50, and the brightness uniformity value threshold is 1.5.

[0060] When obtaining the position characteristics of the light spot, it includes obtaining the coordinates of the centroid of each light spot in the preprocessed image in the image coordinate system and calculating the distance between the centroids of adjacent light spots; the image coordinate system is established with any pixel point of the preprocessed image as the origin, the second direction as the x-axis direction, and the first direction as the y-axis direction. The directions of the x-axis and y-axis in the image coordinate system are Figure 1 The directions of the x-axis and y-axis shown in are the same. The spacing between the centroids of adjacent light spots is determined by the spacing between adjacent line lasers and the spacing between adjacent wafers. Therefore, for an effective light spot, the spacing between the centroids of adjacent light spots should be within a preset range, i.e., the effective light spot spacing range. The centroid coordinates can be determined by the grayscale centroid method or the like.

[0061] According to the above description, it can be known that the edge of each wafer will reflect multiple light spots under the irradiation of multiple beam lasers. Ideally, the centroids of the multiple light spots reflected from the edge of the same wafer have the same y-axis coordinate value in the image coordinate system, so the light spots reflected by each wafer can be divided by the y-axis coordinate value. However, in actual detection, errors caused by reasons such as image acquisition will cause the y-axis coordinates of the centroids of the multiple light spots reflected from the edge of the same wafer to float within a certain error range and will not be exactly the same, but by limiting the y-axis coordinate range, the light spots reflected by each wafer can still be divided by the y-axis coordinate value. Therefore, in this embodiment, multiple y-axis coordinate threshold ranges are set to divide the light spots reflected by each wafer, that is, the y-axis coordinate threshold range corresponds one-to-one to the wafers in the wafer group.

[0062] Since the wafer is vertically arranged along the y-axis direction, the multiple light spots reflected by each wafer should be arranged linearly in the preprocessed image. Therefore, whether the light spot is a valid light spot can be determined by judging whether the light spot centroid coordinates within the same y-axis coordinate threshold range are arranged linearly. The specific judgment method is to divide the light spot centroid coordinates into multiple light spot centroid coordinate groups according to multiple y-axis coordinate threshold ranges, and perform linear fitting on the light spot centroid coordinates in each light spot centroid coordinate group to obtain a fitting straight line; calculate the residual between the light spot centroid coordinates of each light spot in each light spot centroid coordinate group and the fitting straight line. For a valid light spot, the residual between the light spot centroid coordinates and the fitting straight line should not be greater than the residual threshold, that is, the corresponding light spot with a residual greater than the residual threshold is removed. Linear fitting can adopt a linear fitting method such as the least squares method.

[0063] When obtaining the position feature of the light spot, it also includes obtaining the minimum vertical distance between the edge of the light spot and each edge of the preprocessed image, that is, the minimum vertical distance between the edge of the light spot and the four sides of the preprocessed image. For a valid light spot, the minimum vertical distance between the edge of the light spot and each edge of the preprocessed image is not less than the edge threshold. This feature is intended to limit the effective light spot to appear within a preset range of the preprocessed image, rather than at the edge of the image. The edge threshold can be determined according to the shooting range of the camera 3. In this embodiment, the edge threshold is 30 pixels.

[0064] The spots that do not meet the characteristics of the effective spots, that is, invalid spots, are removed from the preprocessed image, and only the effective spots are retained, that is, the effective spots are screened out. After the effective spots are screened out, the features of all the effective spots are extracted to obtain the feature point coordinates of each effective spot in the image coordinate system. Specifically, the edge information in the preprocessed image after removing the invalid spots is extracted by the Canny edge detection algorithm, the edge points of the effective spots are determined, and the connected edge points are marked as a connected domain through the connected domain analysis, and the centroid coordinates of the connected domain are used as the feature point coordinates of the effective spots.

[0065] All feature points are plotted in the image coordinate system to form a wafer group distribution map. When judging the integrity of the wafer group according to the wafer group distribution map, the effective light spot feature points are divided into multiple groups of effective light spot feature point groups according to multiple y-axis coordinate threshold ranges. Each wafer in the wafer group corresponds to each group of effective light spot feature point groups, and the number of effective light spot feature points in each group of effective light spot feature point groups corresponds to the number of light spots reflected by the same wafer. The number of effective light spot feature points in each group of effective light spot feature point groups is compared with the integrity threshold. When the number of effective light spot feature points is less than the integrity threshold, it indicates that a wafer in wafer group 1 is completely missing.

[0066] As an example, Figure 3FIG. 1 shows a side view of 11 line laser beams irradiating the edge of a wafer in wafer group 1, wherein a red straight line represents the line laser and a black straight line represents the wafer. Figure 4 Shows the detection Figure 3 The wafer group distribution diagram obtained for wafer group 1 is shown, Figure 4 The red dots in the figure represent feature points. Figure 3 As shown, the wafer group 1 is composed of 25 wafers, of which the fourth, seventh, thirteenth, seventeenth and twenty-fourth wafers are missing from the wafer blue. Correspondingly, in the wafer distribution diagram, the number of effective light spots in the effective light spot feature point coordinate group corresponding to the fourth, seventh, thirteenth, seventeenth and twenty-fourth wafers is less than the integrity threshold. The integrity threshold is set according to the number of line lasers and can be equal to or slightly less than the number of line lasers.

[0067] It should be noted that Figure 4 This is the distribution diagram of the wafer group under ideal conditions. Generally, the spacing between multiple laser beams is equal, that is, the multiple laser beams are arranged in parallel at equal distances. However, when the side of the wafer in wafer group 1 is actually irradiated, since the side of the wafer is in an arc shape, in the detection image collected by the camera, for the light spots reflected from the edge of the same wafer, the spacing between adjacent light spots in the x-axis direction will first increase and then decrease. In the wafer group distribution diagram formed based on the detection image, as shown in Figure 5 As shown, the spacing between adjacent effective light spot feature points corresponding to the edge of the same wafer in the x-axis direction will first increase and then decrease, but this will not affect the detection effect of the wafer integrity.

[0068] Furthermore, the wafers in the wafer group 1 may also be damaged at the local edges due to collisions during the semiconductor production process. Detecting the integrity of the wafer group 1 also includes detecting the damage to the wafer edge. When the wafer edge is damaged, the number of effective light spots formed by its reflection will also decrease relative to when it is not damaged. Therefore, the wafer group distribution map can also be used to detect whether there is damage to the wafer edge in the wafer group 1. Specifically, a damage threshold greater than the integrity threshold can be set. When the number of feature points in the effective light spot feature point group is less than the damage threshold and greater than the integrity threshold, it can be determined that the corresponding wafer edge is damaged. In addition, in the wafer group distribution map, when the number of feature points within a preset area of ​​a certain area is less than a preset value, that is, when there is a concentrated lack of feature points, it can also be determined that the wafer in the area is damaged. During the detection, multiple wafers in the wafer group remain stationary or rotate synchronously with the wafer center as the rotation center. When multiple wafers rotate synchronously with the wafer center as the rotation center, the damage to the complete edge of the wafer can be detected by rotating the wafer. That is, the wafer edge in the wafer group is divided into multiple sections, and a wafer group distribution map corresponding to a section of the wafer edge is obtained each time the wafer rotates, so as to detect the damage of the complete edge of the wafer according to the wafer group distribution map.

[0069] It should be noted that the words "first" and "second" used in the above description are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. The above is only a preferred embodiment of the present invention, and the present invention is not limited to the above embodiments. It is understood that other improvements and changes directly derived or associated by those skilled in the art without departing from the spirit and concept of the present invention should be considered to be included in the scope of protection of the present invention.

Claims

1. A wafer group integrity detection method based on multi-line laser, characterized in that: The wafer group includes a plurality of wafers arranged vertically and spaced apart along a first direction; When detecting the integrity of the wafer group, a plurality of lasers are used to generate a plurality of line laser beams that are arranged at intervals and parallel to each other along a second direction, and each line laser beam irradiates the edges of all wafers in the wafer group, and the second direction is perpendicular to the first direction; Collecting a detection image, wherein the detection image includes all light spots formed by the reflection of multiple laser beams from the edges of multiple wafers; The detected image is preprocessed to obtain a preprocessed image, and after the preprocessing, the light spot in the preprocessed image is extracted, the light spot feature is obtained, and the effective light spot is screened according to the light spot feature; Feature extraction is performed on all effective light spots to obtain the feature point coordinates of each effective light spot in the image coordinate system, a wafer group distribution map is formed based on the feature point coordinates of all effective light spots, and the integrity of the wafer group is judged according to the wafer group distribution map.

2. The wafer group integrity detection method based on multi-line laser according to claim 1, characterized in that: The image coordinate system is established with any pixel point of the preprocessed image as the origin, the second direction as the x-axis direction, and the first direction as the y-axis direction; The characteristic points of all the effective light spots are plotted in an image coordinate system to form a wafer group distribution map, and the integrity of the wafer group is judged according to the wafer group distribution map, including: Setting a plurality of y-axis coordinate threshold ranges, wherein the y-axis coordinate threshold ranges correspond one to one to the wafers in the wafer group; The effective spot feature points are divided into a plurality of effective spot feature point groups according to a plurality of y-axis coordinate threshold ranges, and the number of effective spot feature points in each effective spot feature point group is compared with the integrity threshold to judge the integrity of the wafer group.

3. The wafer group integrity detection method based on multi-line laser according to claim 2, characterized in that: When obtaining the light spot features, it includes obtaining the geometric features, brightness features and position features of the light spot; When obtaining an effective light spot by screening the light spot characteristics, the light spot whose geometric characteristics, brightness characteristics and position characteristics meet the effective light spot characteristics is taken as the effective light spot; When obtaining the geometric features of the light spot, it includes calculating the area of ​​the light spot, the circularity C, and the aspect ratio of the minimum circumscribed rectangle; the circularity C of the light spot=4πA / P2, where A is the area of ​​the light spot and P is the perimeter of the light spot.

4. The wafer group integrity detection method based on multi-line laser according to claim 3 is characterized in that: When preprocessing the detection image, it includes graying the detection image; When obtaining the brightness characteristics of the light spot, it includes obtaining the average gray value, contrast and brightness uniformity of the light spot, among which, The average grayscale value of the light spot is the average grayscale value of all pixels in the light spot, the contrast of the light spot is the difference between the average grayscale value of the light spot and the average grayscale value of the background in the preprocessed image, and the uniform brightness value of the light spot is the ratio of the maximum grayscale value to the minimum grayscale value of the pixels in the light spot.

5. The wafer group integrity detection method based on multi-line laser according to claim 3, characterized in that: When obtaining the position characteristics of the light spot, it includes: Obtain the minimum vertical distance between the edge of the light spot and each edge of the preprocessed image; Obtain the coordinates of the centroid of each light spot in the preprocessed image in the image coordinate system, and calculate the distance between the centroids of adjacent light spots; Dividing the spot centroid coordinates into a plurality of spot centroid coordinate groups according to a plurality of y-axis coordinate threshold ranges, and performing straight line fitting on the spot centroid coordinates in each spot centroid coordinate group to obtain a fitting straight line; The residual between the centroid coordinates of each spot in each spot centroid coordinate group and the fitting straight line is calculated.

6. The wafer group integrity detection method based on multi-line laser according to claim 4, characterized in that: When the geometric features meet the effective light spot features, the features include: the area of ​​the light spot is within the effective light spot area range, the circularity C of the light spot is not less than the circularity threshold, and the aspect ratio of the minimum circumscribed rectangle of the light spot is within the effective light spot aspect ratio range; When the brightness characteristics meet the effective light spot characteristics, they include: the average gray value of the light spot is not less than the average gray value threshold, the contrast of the light spot is not less than the contrast threshold, and the brightness uniformity of the light spot is not greater than the brightness uniformity threshold.

7. The wafer group integrity detection method based on multi-line laser according to claim 5, characterized in that: When the position characteristics meet the effective light spot characteristics, they include: The spacing between adjacent spot centroids is within the effective spot spacing range, the minimum vertical distance between the spot edge and each edge of the preprocessed image is not less than the edge threshold, and the residual between the spot centroid coordinates and the corresponding fitting straight line is not greater than the residual threshold.

8. The wafer group integrity detection method based on multi-line laser according to claim 1, characterized in that: When extracting features from any effective light spot, it includes determining the edge points of the effective light spot through an edge detection algorithm, marking the connected edge points as a connected domain through connected domain analysis, and using the centroid coordinates of the connected domain as the feature point coordinates of the effective light spot.

9. The wafer group integrity detection method based on multi-line laser according to claim 4, characterized in that: The preprocessing of the detection image also includes performing noise reduction processing on the grayscale detection image, and the noise reduction processing at least includes performing morphological filtering on the grayscale detection image.

10. The wafer group integrity detection method based on multi-line laser according to claim 4, characterized in that: The centers of multiple wafers in the wafer group coincide with each other. During the inspection, the multiple wafers in the wafer group remain stationary or rotate with the center of the wafer as the rotation center.

Images