Edge recognition method, device, computer equipment and wafer laser cutting method

Abstract

The present application relates to an edge recognition method, device, computer equipment and wafer laser cutting method. The method includes: acquiring a groove image; constructing a rotating rectangle on the groove image, and determining the projection direction of the rotating rectangle; the projection direction is parallel to the direction of wafer laser grooving; determining each sampling point in the rotating rectangle along the projection direction, collecting the grayscale value of the pixel point corresponding to each sampling point, and generating a projection grayscale curve; obtaining the extreme value points on the projection grayscale curve by derivative according to the projection grayscale curve; screening the extreme value points, and determining the search range according to the screened extreme value points; identifying the groove edge within the search range. Under different materials and different laser grooving conditions, the edge recognition method is less dependent on environmental factors, avoiding edge recognition failure or misidentification problems caused by different characteristics of wafer materials, and coping with real-time edge recognition tasks under different materials and different laser grooving conditions.

Description

Technical Field

[0001] The present application relates to the field of semiconductor manufacturing technology, and in particular to an edge recognition method, device, computer equipment and wafer laser cutting method. Background Art

[0002] With the increasing requirements for the integration and performance of integrated circuits, the precision requirements for trenching the interlayer dielectric on the wafer substrate are becoming higher and higher.

[0003] At present, although the commonly used wafer laser grooving method can meet the precision requirements for wafer grooving, during the laser cutting process, there will be slag splashing, wafer material turning inside out and other dirty and blurred wafer groove edges. In addition, due to factors such as the diversity of wafer materials, the large number of product types, and the complexity of wafer groove images, it is difficult to accurately identify the cut marks of laser grooving, making it impossible to perform real-time detection of wafer grooves in a timely manner.

[0004] In traditional technology, it is often necessary to judge the processing effect and processing accuracy of laser grooving only after the cutting is completed and impurities are cleaned. It is impossible to compensate the accuracy of laser cutting in time, which leads to errors between the actual groove and the designed groove due to system errors in laser grooving, ultimately resulting in poor product quality. Summary of the invention

[0005] Based on this, it is necessary to provide an edge recognition method, device, computer equipment and wafer laser cutting method that can identify the groove position in real time during the wafer laser grooving process to address the above technical problems.

[0006] In a first aspect, the present application provides an edge recognition method. The method comprises:

[0007] Acquire a groove image; wherein the groove image is obtained by photographing the groove of the wafer after process processing;

[0008] Construct a rotating rectangle on the groove image and determine the projection direction of the rotating rectangle; the projection direction is parallel to the direction of laser grooving of the wafer;

[0009] In the groove image, each sampling point in the rotating rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate a projection grayscale curve;

[0010] Determine the extreme points of the projection grayscale curve, and determine the groove edge of the wafer based on the extreme points; obtain the extreme points on the projection grayscale curve by taking the derivative of the projection grayscale curve; screen the extreme points, and determine the search range based on the screened extreme points; the search range is the area where the groove edge of the wafer is located; identify the groove edge of the wafer within the search range.

[0011] In one embodiment, a rotating rectangle is constructed on the groove image, and a projection direction of the rotating rectangle is determined; after the projection direction is parallel to the direction of laser grooving of the wafer, the method includes:

[0012] Determines the search direction of the rotated rectangle; the search direction is perpendicular to the projection direction and is used to identify the edge of the groove along the search direction.

[0013] In one embodiment, the method further comprises:

[0014] A plurality of rotating rectangles are arranged along the laser grooving direction of the wafer;

[0015] Obtain the projection grayscale curves of multiple rotated rectangles respectively, and obtain the extreme value points on each projection grayscale curve;

[0016] The trench edge of the wafer is determined based on the connecting line between the extreme points.

[0017] In one embodiment, the extreme value points are screened, and the search range is determined according to the screened extreme value points, including:

[0018] When there is only one maximum point on the projected grayscale curve but no minimum point, the first-order derivative and the second-order derivative of the projected grayscale curve are performed to obtain the maximum point of the first-order derivative of the projected grayscale curve and the minimum point determined by the second-order derivative, and the interval between the minimum point of the second-order derivative of the projected grayscale curve and the maximum point of the first-order derivative of the projected grayscale curve is used as the search range;

[0019] In the case where there is a minimum point on the projected grayscale curve, when there is a maximum value, the interval between the maximum point and the minimum point of the projected grayscale curve is the search range; when there is no maximum value, the interval between the minimum point of the projected grayscale curve and the edge of the groove is the search range;

[0020] When there are multiple minimum points on the projected grayscale curve, the minimum point among the minimum points is obtained according to the projected grayscale curve, and the adjacent maximum point is calculated, with the interval between the minimum point and the maximum point of the projected grayscale curve as the search range.

[0021] In one embodiment, in the groove image, each sampling point in the rotating rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate the projection grayscale curve, which includes:

[0022] A plurality of groups of sampling points are set in the groove image along the projection direction, and each group of sampling points is evenly spaced in the search direction;

[0023] Sequentially obtain the four pixels closest to each sampling point, and obtain the grayscale value of the sampling point through distance interpolation;

[0024] The projection grayscale mean of each group of sampling points in the horizontal projection direction is obtained, and a projection grayscale curve is constructed according to the projection grayscale mean.

[0025] In one embodiment, the four pixels closest to each sampling point are sequentially obtained, and the grayscale value of the sampling point is obtained by distance interpolation, which includes:

[0026] Filter the maximum grayscale value and the minimum grayscale value in the rotated rectangle;

[0027] The grayscale data within the rotated rectangle is normalized according to the maximum grayscale value and the minimum grayscale value.

[0028] In one embodiment, the method further comprises:

[0029] A mask area is created according to the position of the wafer trench; the mask area is used to cover the area away from the edge of the wafer trench;

[0030] Perform image grayscale stretching on non-mask areas.

[0031] In a second aspect, the present application also provides a wafer laser cutting method. The method comprises:

[0032] Laser grooving the wafer according to a preset wafer laser grooving direction;

[0033] According to the image detection method of any embodiment of the first aspect, the edge of the groove after the wafer laser groove is identified to determine whether the edge of the groove is consistent with the preset wafer laser groove direction;

[0034] When the groove edge is inconsistent with the preset wafer laser grooving direction, the preset wafer laser grooving direction is accurately compensated.

[0035] In a third aspect, the present application also provides an edge recognition device. The device comprises:

[0036] An image acquisition module is used to obtain a groove image; wherein the groove image is obtained by photographing the groove of the wafer after the process treatment;

[0037] The initial module is used to construct a rotating rectangle on the groove image and determine the projection direction of the rotating rectangle; the projection direction is parallel to the direction of laser grooving of the wafer;

[0038] A sampling module is used to determine each sampling point in the rotating rectangle along the projection direction in the groove image, and collect the gray value of the pixel point corresponding to each sampling point to generate a projection gray curve;

[0039] A search module is used to determine the extreme points of the projection grayscale curve, and determine the groove edge of the wafer based on the extreme points; obtain the extreme points on the projection grayscale curve by deriving the projection grayscale curve; screen the extreme points, and determine the search range based on the screened extreme points; the search range is the area where the groove edge of the wafer is located; and identify the groove edge of the wafer within the search range.

[0040] In a fourth aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0041] Acquire a groove image; wherein the groove image is obtained by photographing the groove of the wafer after process processing;

[0042] Construct a rotating rectangle on the groove image and determine the projection direction of the rotating rectangle; the projection direction is parallel to the direction of laser grooving of the wafer;

[0043] In the groove image, each sampling point in the rotating rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate a projection grayscale curve;

[0044] Determine the extreme points of the projection grayscale curve, and determine the groove edge of the wafer based on the extreme points; obtain the extreme points on the projection grayscale curve by taking the derivative of the projection grayscale curve; screen the extreme points, and determine the search range based on the screened extreme points; the search range is the area where the groove edge of the wafer is located; identify the groove edge of the wafer within the search range.

[0045] The above-mentioned edge recognition method, device, computer equipment and wafer laser cutting method maintain high-precision edge recognition capability by constructing a rotating rectangle on the groove image and performing grayscale projection analysis when the wafer material has low sensitivity to the light source and the groove edge is blurred. By analyzing the extreme points of the projection grayscale curve, the extreme points on the projection grayscale curve are obtained. The extreme points on the projection grayscale curve represent the points where the grayscale change of the groove image is most drastic, corresponding to the edge position of the wafer groove, and then the edge position of the groove is obtained. Based on this, under different materials and different laser grooving conditions, the edge recognition method is less dependent on environmental factors, maintains good robustness and adaptability, avoids edge recognition failure or misidentification problems caused by different characteristics of wafer materials, and can cope with real-time edge recognition tasks under different materials and different laser grooving conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A diagram showing an application environment of an edge recognition method in an embodiment;

[0047] Figure 2 is a schematic flow chart of an edge recognition method in one embodiment;

[0048] Figure 3 A schematic diagram of the positions of a rotating rectangle and sampling points in one embodiment;

[0049] Figure 4 is a schematic diagram of a first-order derivative function curve of a projected grayscale curve in one embodiment;

[0050] Figure 5 It is a schematic diagram of a first-order derivative function curve of a projection grayscale curve in another embodiment;

[0051] Figure 6 It is a schematic diagram of a first-order derivative function curve of a projection grayscale curve in another embodiment;

[0052] FIG7( a ) is a grayscale image of a wafer groove before normalization processing in one embodiment;

[0053] FIG7( b ) is a normalized grayscale image of a wafer groove in one embodiment;

[0054] Figure 8 is a schematic diagram of a mask area on a groove image in one embodiment;

[0055] Fig. 9 is a schematic flow chart of an edge recognition method in another embodiment;

[0056] Fig.10 A schematic diagram of a process of a wafer laser cutting method in one embodiment;

[0057] Fig.11 is a schematic diagram of the structure of an edge recognition device in one embodiment;

[0058] Fig.12 A schematic diagram of the structure of a wafer laser cutting device in one embodiment;

[0059] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0061] The edge recognition method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. (Described in conjunction with the overall solution of right 1). Among them, the terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0062] In one embodiment, the first aspect is as follows: Figure 2 As shown, an edge recognition method is provided, which is applied to the groove edge recognition of wafer laser grooving. Figure 1 The server in the example is used to illustrate the following steps:

[0063] Step 202, obtaining a groove image; wherein the groove image is obtained by photographing the grooves of the wafer after the process.

[0064] The groove image is an image of the groove structure formed on the surface of the wafer after laser cutting. There are tiny cutting marks or spatters on the edge of the groove structure, and there may be slag deposition in the cutting path.

[0065] Specifically, a high-resolution camera is used to photograph the wafer after laser grooving. Since the wafer material is insensitive to the light source, the light absorption rate is low and the surface will cause light scattering or diffuse reflection after exposure to light, the groove edge is not easy to be identified under light sources of different intensities; therefore, the captured groove image should maintain color accuracy and image accuracy as much as possible to facilitate the subsequent capture of grayscale information.

[0066] Step 204, construct a rotated rectangle on the groove image, and determine the projection direction of the rotated rectangle; the projection direction is parallel to the direction of laser grooving of the wafer.

[0067] Among them, the rotated rectangle is a rectangle used to analyze a specific area of ​​an image. It is often used for grayscale analysis and shape matching of local areas in image processing. It can be rotated around the center point at any angle to adapt to specific image features or object edges. The direction of wafer laser grooving refers to the trajectory of the laser moving on the surface of the wafer. In the coordinate system of the wafer, the direction of the laser grooving should be parallel or perpendicular to a certain reference line of the wafer. The projection direction refers to the direction in which the grayscale values ​​of pixels in the image are summarized along a specific direction. For example, the grayscale values ​​can be accumulated or averaged along the horizontal or vertical direction.

[0068] Specifically, the position of the groove edge is preliminarily estimated according to the direction and position of the laser grooving, and according to the expected width of the groove, a rotating rectangle is constructed on the acquired groove image, the aspect ratio of the rotating rectangle is set, and the projection direction of the rotating rectangle is determined. Optionally, the rotating rectangle can cover only one side edge of the groove, or both sides of the groove; the approximate edge of the groove can be used as the center of the rotating rectangle, or the midline of the groove can be used as the center of the rotating rectangle.

[0069] Step 206 , in the groove image, each sampling point in the rotation rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate a projection grayscale curve.

[0070] Specifically, multiple sampling points are evenly distributed along the projection direction inside the rotating rectangle. Taking the projection direction as the reference, the sampling points on each row parallel to the projection direction are a group of sampling points. The interval between the sampling points is determined according to the aspect ratio of the rotating rectangle and the resolution of the groove image, and each group of sampling points is evenly distributed to ensure that the entire rotating rectangle can be covered. The grayscale value of each pixel is obtained, and the grayscale mean of each group in the projection direction is determined according to the grayscale value of each pixel, thereby generating a projection grayscale curve.

[0071] Step 208, determine the extreme points of the projection grayscale curve, and determine the groove edge of the wafer based on the extreme points; obtain the extreme points on the projection grayscale curve by derivative according to the projection grayscale curve; filter the extreme points, and determine the search range based on the filtered extreme points; the search range is the area where the groove edge of the wafer is located; identify the groove edge of the wafer within the search range.

[0072] Specifically, at the edge of an object in an image, the grayscale value will change significantly. For example, the area on one side of the edge may be bright, while the area on the other side may be dark. This drastic grayscale change appears as a local extreme point on the projected grayscale curve. Therefore, the extreme point on the projected grayscale curve is obtained to determine the edge position of the wafer groove.

[0073] In the above edge recognition method, when the wafer material has low sensitivity to the light source and the groove edge is blurred, a high-precision edge recognition capability is maintained by constructing a rotated rectangle on the groove image and performing grayscale projection analysis. The extreme points on the projection grayscale curve are obtained by analyzing the extreme points of the projection grayscale curve. The extreme points on the projection grayscale curve represent the points where the grayscale change of the groove image is most drastic, corresponding to the edge position of the wafer groove, and then the edge position of the groove is obtained. Based on this, under different materials and different laser grooving conditions, the edge recognition method is less dependent on environmental factors, maintains good robustness and adaptability, avoids edge recognition failures or misidentification problems caused by different characteristics of wafer materials, and can cope with edge recognition tasks under different materials and different laser grooving conditions.

[0074] See also Figure 3 In one embodiment, a rotating rectangle is constructed on the groove image, and the projection direction of the rotating rectangle is determined; after the projection direction is parallel to the direction of laser grooving of the wafer, it includes: determining the search direction of the rotating rectangle; the search direction is perpendicular to the projection direction, and is used to identify the edge of the groove along the search direction.

[0075] Among them, the search direction is the scanning direction used to locate or identify the target features. During the edge recognition process, the search direction is perpendicular to the projection direction and is used to scan in a direction perpendicular to the laser cutting path to extract the image feature information of the relevant wafer grooves to determine the edge position of the grooves.

[0076] In this embodiment, a search direction perpendicular to the projection direction is set. By scanning in the search direction, the significant features of the edge in the image can be captured more clearly, and the problem of inaccurate edge recognition caused by unclear grayscale changes or noise interference in the projection direction can be effectively avoided. During the laser grooving process, due to process and wafer material problems, the groove edge may appear irregular or slightly distorted, and traditional edge recognition methods may have difficulty in accurately identifying such edges; by scanning in the search direction, it is possible to better adapt to the changes in these edges, enhance the robustness of the edge recognition method, and enhance the stability and reliability under different process conditions.

[0077] In one embodiment, determining the extreme points of the projection grayscale curve, and determining the groove edge of the wafer based on the extreme points includes: obtaining the extreme points on the projection grayscale curve by derivative according to the projection grayscale curve; screening the extreme points, and determining the search range based on the screened extreme points; the search range is the area where the groove edge is located; and identifying the groove edge within the search range.

[0078] The projected grayscale curve is a grayscale curve obtained by accumulating or averaging the grayscale values ​​of the sampling points in the projection direction of the rotating rectangle; the projected grayscale curve reflects the change in brightness or grayscale in the search direction of the groove image. The extreme point is the point in the projected grayscale curve where the grayscale value changes dramatically. Because the edge of the object is usually the place where the grayscale changes most dramatically, the extreme point usually represents the location of the edge in the groove image. The search range is distributed along the search range and is determined by the extreme point; the search range is usually within a certain range around the extreme point or between multiple extreme points.

[0079] Exemplarily, the laser cutting path is a straight line and a preset path, while the groove of the wafer is the actual cutting path. After laser cutting, the grayscale value at the edge of the groove will change dramatically, which is reflected as an extreme point on the projected grayscale curve. The search range is determined according to the extreme point, and the groove edge is determined along the search direction. It should be understood that when there is a large deviation between the extreme point and the laser cutting path, it indicates that there is a deviation between the edge of the wafer groove and the laser cutting path.

[0080] Optionally, multiple rotating rectangles are provided on the laser cutting path, and the projection grayscale curves of the multiple rotating rectangles are obtained respectively to obtain the extreme points on each projection grayscale curve. The edge position of the groove is determined according to the position of the extreme points on each projection grayscale curve on the wafer coordinates, and the line between the extreme points should be the edge position of the wafer groove.

[0081] In this embodiment, by taking the derivative of the projected grayscale curve to obtain the extreme point, identifying the extreme point, the edge position of the groove can be effectively determined; at the same time, the search range is determined based on the area around the extreme point, which reduces the search area, avoids misidentification or noise interference, and improves the accuracy and reliability of edge identification. According to these extreme points, the edge position of the wafer groove can be accurately located, the accuracy and stability of edge identification can be improved, and the identification result can be ensured to be more accurate, which is suitable for different wafer cutting processes.

[0082] See also Figures 4 to 6 In a further embodiment, the extreme points are screened, and the search range is determined according to the screened extreme points, including: when there is only one maximum point on the projected grayscale curve but no minimum point, the first-order derivative and the second-order derivative of the projected grayscale curve are taken to obtain the maximum point of the first-order derivative of the projected grayscale curve and the minimum point determined by the second-order derivative, and the interval between the minimum point of the second-order derivative of the projected grayscale curve and the maximum point of the first-order derivative of the projected grayscale curve is used as the search range; when there is one minimum point on the projected grayscale curve, when there is a maximum value, the interval between the maximum point and the minimum point of the projected grayscale curve is used as the search range; when there is no maximum value, the interval between the minimum point of the projected grayscale curve and the edge of the groove is used as the search range; when there are multiple minimum points on the projected grayscale curve, the minimum point among the minimum points is obtained according to the projected grayscale curve, and the adjacent maximum point is calculated, and the interval between the minimum point and the maximum point of the projected grayscale curve is used as the search range.

[0083] Specifically, the projection grayscale curve is derived to obtain the extreme point of the grayscale projection curve. Figure 4, take the first-order derivative and the second-order derivative of the grayscale projection curve, find the maximum point x0 of the first-order derivative, find the minimum point x1 of the second-order derivative, and x1>x0, x1 is the point where the gradient of the groove edge changes most strongly, the starting point of the search range is x1, and the ending point is x0.

[0084] Please combine Figure 5 When there is only one minimum point on the projected grayscale curve, take the first-order derivative of the grayscale projection curve to obtain the maximum point x0 and the minimum point x1 of the projected grayscale curve, and x1>x0, x1 is the minimum grayscale point of the groove edge, then the starting point of the search range is x1 and the ending point is x0.

[0085] Please combine Figure 6 When there are multiple minimum points on the projected grayscale curve, take the first-order derivative of the grayscale projection curve to obtain the maximum point x0 and the minimum point x1 of the projected grayscale curve, and x1>x0, x1 is the minimum grayscale point of the groove edge, then the starting point of the search range is x1 and the ending point is x0.

[0086] In this embodiment, by analyzing the extreme points and their changing trends, a reasonable search range is determined, thereby accurately locating the edge of the groove. When the projected grayscale curve is relatively simple, the combination of the second-order derivative and the first-order derivative can accurately define the search area; and when the projected grayscale curve is relatively complex, the edge recognition process is further optimized by selecting the minimum point with the smallest grayscale as the starting point of the search range. This method reduces noise interference and ensures efficient and accurate edge detection.

[0087] In one embodiment, in the groove image, each sampling point in the rotated rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected, and the generation of the projection grayscale curve includes: setting multiple groups of sampling points along the projection direction in the groove image, and each group of sampling points is evenly spaced in the search direction; sequentially obtaining the four pixels closest to each sampling point, and obtaining the grayscale value of the sampling point by distance interpolation; obtaining the projection grayscale mean of each group of sampling points in the horizontal projection direction, and constructing the projection grayscale curve according to the projection grayscale mean.

[0088] Specific, combined Figure 3 As shown, multiple groups of sampling points are evenly set along the projection direction, each group of sampling points is evenly spaced in the search direction, and each sampling point is evenly arranged in the projection direction and the search direction to ensure that the entire rotating rectangular area can be covered. According to the position of each sampling point, the four pixels closest to the sampling point are obtained in turn, and the grayscale value of each sampling point is obtained by distance interpolation. Optionally, the grayscale value of each sampling point is obtained by bilinear interpolation, nearest neighbor difference method, bicubic interpolation or Lagrange interpolation; the grayscale value of each sampling point is obtained in turn along the projection direction.

[0089] Taking bilinear interpolation as an example, the grayscale value of each set of sampling points is collected:

[0090]

[0091] In the formula, gray n is the gray value of each sampling point, g i is a set of grayscale values ​​of a set of sampling points. Interpolate all sampling points, find the four pixels closest to the sampling points, get the grayscale values ​​of the sampling points according to the bilinear interpolation method, and get the mean of the horizontal grayscale projection of each set of sampling points:

[0092]

[0093] In the rotated rectangle, the data set of the mean horizontal grayscale projection of each set of sampling points is:

[0094] y={y0 ,y1, …, y n-1}

[0095] The projected grayscale curve is obtained by fitting the polynomial grayscale value curve by least squares and Gaussian smoothing the curve:

[0096] y=a0+a1x+a2x 2 +…+a k x k

[0097] Derive the projected grayscale curve:

[0098]

[0099] For a1, a2,…, a k Take the derivative and set the first-order derivative of the projected grayscale curve to 0, and get n equations:

[0100]

[0101]

[0102] …

[0103]

[0104] Simplifying and rearranging, we can get the following matrix equation:

[0105]

[0106] We can get C = , we can obtain a1, a2,…, a respectively. k , and then get the projection grayscale curve equation y.

[0107] In this embodiment, the grayscale value of the sampling point is obtained according to the distance interpolation method, and then a projected grayscale curve is generated, which can accurately reflect the grayscale changes of the groove image in the search direction, especially in the edge area, and effectively support subsequent edge recognition; through precise interpolation and grayscale calculation, the edge recognition accuracy can be improved, the error and noise interference can be reduced, and the image analysis effect can be optimized.

[0108] In one embodiment, the four pixels closest to each sampling point are sequentially obtained, and the grayscale value of the sampling point is obtained by distance interpolation, which includes: screening the maximum grayscale value and the minimum grayscale value in the rotated rectangle; and normalizing the grayscale data in the rotated rectangle according to the maximum grayscale value and the minimum grayscale value.

[0109] Among them, normalization processing refers to converting data into a specific range or standard in some way, so that different values ​​have the same measurement scale, which is convenient for comparison or processing.

[0110] Combined with Figure 7 (a) and Figure 7 (b), specifically, because the wafer material is insensitive to the light source, the image quality of the captured groove image is poor, and then the grayscale projection data is normalized to the range of [0-255]. First, the grayscale value of each pixel in the rotated rectangle is counted, and the maximum grayscale value and the minimum grayscale value in the rotated rectangle are screened out. All the pixels in the rotated rectangle are normalized by the following formula to obtain the grayscale value of the processed pixel:

[0111]

[0112] In this embodiment, the grayscale data in the rotated rectangle is normalized according to the maximum grayscale value and the minimum grayscale value. Through normalization, the influence of different grayscale value ranges can be eliminated, the consistency and comparability of the data can be ensured, and the problem of uneven grayscale or noise interference in the subsequent processing can be avoided. The normalized data makes the subsequent analysis more accurate, especially when performing edge recognition, it can more accurately reflect the edge features in the image and improve the accuracy and robustness of recognition.

[0113] In one embodiment, the method further includes: creating a mask area according to the position of the wafer groove; the mask area is used to shield the area away from the edge of the wafer groove; and performing image grayscale stretching on the non-mask area.

[0114] The mask region refers to the image region that is used to mark or block the image region that does not need to be processed or paid attention to during the image processing process; the mask is usually a binary image that represents the area that needs to be processed and the area that does not need to be processed.

[0115] Exemplarily, the mask of the present application is an image area that does not need to be paid attention to, that is, an area far away from the edge of the wafer groove. According to the laser cutting path information, multiple mask areas are defined in the groove image, and the area includes the surrounding part of the wafer groove, such as Figure 8 As shown, Mask1, Mask2 and Mask3 are mask areas. The grayscale of the non-mask area is stretched to enhance the contrast of the non-groove area in the image and adjust the grayscale level to make the details more prominent.

[0116] In this embodiment, by creating a mask area, the groove edges in the groove image can be accurately processed and concentrated, avoiding processing of irrelevant background areas and grooves, thereby improving the efficiency and accuracy of edge recognition. Grayscale stretching of the non-mask area can enhance the contrast of the image, so that the areas with low grayscale values ​​are fully enhanced, and the image details are improved; in particular, the grayscale difference of the groove edge is more obvious, which helps to accurately identify the groove edge and improve the accuracy of subsequent edge detection.

[0117] In one embodiment, Fig. 9 As shown, an edge recognition method is provided, the method comprising:

[0118] Step 902, obtaining a groove image; wherein the groove image is obtained by photographing the grooves of the wafer after the process.

[0119] Step 904, creating a mask area according to the position of the wafer trench; the mask area is used to shield the area away from the edge of the wafer trench.

[0120] Step 906: perform grayscale stretching on the non-mask area.

[0121] Step 908, construct a rotated rectangle on the groove image, and determine the projection direction of the rotated rectangle; the projection direction is parallel to the direction of laser grooving of the wafer.

[0122] Step 910, determining the search direction of the rotated rectangle; the search direction is perpendicular to the projection direction and is used to identify the edge of the groove along the search direction.

[0123] Step 912: multiple groups of sampling points are set in the groove image along the projection direction, and each group of sampling points is evenly spaced in the search direction.

[0124] Step 914, sequentially obtain the four pixels closest to each sampling point, and obtain the grayscale value of the sampling point by distance interpolation.

[0125] Step 916, screening the maximum grayscale value and the minimum grayscale value in the rotated rectangle.

[0126] Step 918, normalizing the grayscale data within the rotated rectangle according to the maximum grayscale value and the minimum grayscale value.

[0127] Step 920, obtaining the projection grayscale mean value of each group of sampling points in the horizontal projection direction, and constructing a projection grayscale curve according to the projection grayscale mean value.

[0128] Step 922, deriving the projected grayscale curve to obtain the extreme value points on the projected grayscale curve.

[0129] Step 924, screen the extreme points, and determine the search range based on the screened extreme points; the search range is the area where the groove edge is located.

[0130] Step 926, when there is only one maximum point but no minimum point on the projected grayscale curve, perform first-order derivative and second-order derivative on the projected grayscale curve to obtain the maximum point of the first-order derivative of the projected grayscale curve and the minimum point determined by the second-order derivative, and take the interval between the minimum point of the second-order derivative of the projected grayscale curve and the maximum point of the first-order derivative of the projected grayscale curve as the search range.

[0131] Step 928, when there is a minimum point on the projected grayscale curve, when there is a maximum value, the interval between the maximum point and the minimum point of the projected grayscale curve is the search range; when there is no maximum value, the interval between the minimum point of the projected grayscale curve and the groove edge is the search range.

[0132] Step 930, when there are multiple minimum points on the projected grayscale curve, obtain the minimum point among the minimum points according to the projected grayscale curve, calculate the adjacent maximum point, and use the interval between the minimum point and the maximum point of the projected grayscale curve as the search range.

[0133] Step 932, identifying the groove edge within the search range.

[0134] In this embodiment, when the wafer material has low sensitivity to the light source and the groove edge is blurred, a high-precision edge recognition capability is maintained by constructing a rotating rectangle on the groove image and performing grayscale projection analysis. By analyzing the extreme points of the projection grayscale curve, the extreme points on the projection grayscale curve are obtained. The extreme points on the projection grayscale curve represent the points where the grayscale change of the groove image is most drastic, corresponding to the edge position of the wafer groove, and then the edge position of the groove is obtained. Based on this, under different materials and different laser grooving conditions, the edge recognition method is less dependent on environmental factors, maintains good robustness and adaptability, avoids edge recognition failures or misidentification problems caused by different characteristics of wafer materials, and can cope with edge recognition tasks under different materials and different laser grooving conditions.

[0135] In a second aspect, in one embodiment, Fig.10 As shown, a wafer laser cutting method is provided, the method comprising:

[0136] Step 1002 , laser grooving the wafer according to a preset wafer laser grooving direction.

[0137] Step 1004, identifying the edge of the groove after the wafer laser grooving according to the edge recognition method of any embodiment of the first aspect, and determining whether the groove edge is consistent with a preset wafer laser grooving direction.

[0138] Step 1006 , when the groove edge is inconsistent with the preset wafer laser grooving direction, precision compensation is performed on the preset wafer laser grooving direction.

[0139] Specifically, the wafer is laser grooved according to the pre-designed laser wafer groove direction, and during the groove process, the groove image of the grooved area is obtained. The groove image is masked, normalized, and grayscale stretched; a rotated rectangle is constructed in the processed area, the grayscale values ​​of the sampling points in the rotated rectangle are sampled, and the average grayscale value in the projection direction is obtained, and a polynomial grayscale curve is fitted to determine the extreme point of the projection grayscale curve, and the groove edge of the wafer is determined based on the extreme point.

[0140] Exemplarily, multiple rotated rectangles are constructed on the groove image, and the extreme points of the projected grayscale curves on each rotated rectangle are obtained to obtain the groove edge points of the wafer on the wafer coordinates. A straight line is fitted to the groove edge points to determine whether the straight line is consistent with the preset wafer laser slotting direction. If they are consistent, it is determined that there is no deviation in the wafer laser slotting direction, or the deviation is within the error range; if they are inconsistent or the deviation is large, the preset wafer laser slotting direction is compensated for accuracy and the laser path is corrected.

[0141] In this embodiment, through the edge recognition method, when the edge of the wafer groove is contaminated and blurred, the position of the groove edge is accurately judged, and then it is judged whether the actually opened groove is the same as the preset wafer laser grooving direction, and then the laser grooving direction is accurately compensated to improve the accuracy and quality of the wafer grooving process; avoid the cutting deviation in the wafer cutting path due to system error accuracy, which causes the reduction of cutting process quality.

[0142] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are shown in sequence, these steps are not necessarily executed in sequence in the order indicated. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0143] In the third aspect, based on the same inventive concept, the embodiment of the present application also provides an edge recognition device for implementing the edge recognition method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more edge recognition device embodiments provided below can refer to the limitations on the image detection method above, and will not be repeated here.

[0144] In one embodiment, Fig.11 As shown, an edge recognition device is provided, including: an image acquisition module 1102, an initial module 1104, a sampling module 1106 and a search module 1108, wherein:

[0145] The image acquisition module 1102 is used to obtain a groove image; wherein the groove image is obtained by photographing the grooves of the wafer after the process treatment.

[0146] The initial module 1104 is used to construct a rotated rectangle on the groove image and determine the projection direction of the rotated rectangle; the projection direction is parallel to the direction of laser grooving of the wafer.

[0147] The sampling module 1106 is used to determine each sampling point in the rotating rectangle along the projection direction in the groove image, and collect the gray value of the pixel point corresponding to each sampling point to generate a projection gray curve.

[0148] The search module 1108 is used to determine the extreme points of the projection grayscale curve, and determine the groove edge of the wafer based on the extreme points; obtain the extreme points on the projection grayscale curve by deriving the projection grayscale curve; filter the extreme points, and determine the search range based on the filtered extreme points; the search range is the area where the groove edge of the wafer is located; and identify the groove edge of the wafer within the search range.

[0149] In one embodiment, the initialization module 1104 further includes:

[0150] The search direction module is used to determine the search direction of the rotating rectangle; the search direction is perpendicular to the projection direction and is used to identify the edge of the groove along the search direction.

[0151] In one embodiment, the apparatus further comprises:

[0152] Repeating the building block is used to set a plurality of rotating rectangles along the laser grooving direction of the wafer.

[0153] The extreme value module is used to obtain the projection grayscale curves of multiple rotated rectangles respectively, and obtain the extreme value points on each projection grayscale curve.

[0154] The connection identification module is used to determine the groove edge of the wafer according to the connection line between the extreme value points.

[0155] In one embodiment, the screening module includes:

[0156] The first screening submodule is used to perform first-order derivative and second-order derivative on the projected grayscale curve when there is only one maximum point but no minimum point on the projected grayscale curve, so as to obtain the maximum point of the first-order derivative of the projected grayscale curve and the minimum point determined by the second-order derivative, and the interval between the minimum point of the second-order derivative of the projected grayscale curve and the maximum point of the first-order derivative of the projected grayscale curve is used as the search range.

[0157] The second screening submodule is used for, when there is only one minimum point on the projected grayscale curve, when there is a maximum value, to use the interval between the maximum point and the minimum point of the projected grayscale curve as the search range; when there is no maximum value, to use the interval between the minimum point of the projected grayscale curve and the groove edge as the search range.

[0158] The third screening submodule is used to obtain the minimum point among the minimum points according to the projected grayscale curve when there are multiple minimum points on the projected grayscale curve, calculate the adjacent maximum point, and use the interval between the minimum point and the maximum point of the projected grayscale curve as the search range.

[0159] In one embodiment, the acquisition module 1106 includes:

[0160] The sampling module is used to set a plurality of groups of sampling points in the groove image along a direction parallel to the projection direction, and each group of sampling points is evenly spaced in the search direction.

[0161] The interpolation module is used to sequentially obtain the four pixels closest to each sampling point and obtain the grayscale value of the sampling point through distance interpolation.

[0162] The curve construction module is used to obtain the projection grayscale mean value of each group of sampling points in the horizontal projection direction, and to construct a projection grayscale curve according to the projection grayscale mean value.

[0163] In one embodiment, the acquisition module 1106 further includes:

[0164] Grayscale statistics module, used to filter the maximum and minimum grayscale values ​​within the rotated rectangle.

[0165] The normalization module is used to normalize the grayscale data in the rotating rectangle according to the maximum grayscale value and the minimum grayscale value.

[0166] In one embodiment, it further includes:

[0167] The mask module is used to create a mask area according to the position of the wafer groove; the mask area is used to cover the area away from the edge of the wafer groove.

[0168] The grayscale stretching module is used to stretch the image grayscale in the non-mask area.

[0169] In the fourth aspect, based on the same inventive concept, the embodiment of the present application also provides a wafer laser cutting device for implementing the wafer laser cutting method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more wafer laser cutting device embodiments provided below can refer to the limitations on the wafer laser cutting method above, and will not be repeated here.

[0170] In one embodiment, Fig.12 As shown, the wafer laser cutting device includes:

[0171] The slotting module 1202 is used to perform laser slotting on the wafer according to a preset wafer laser slotting direction.

[0172] The calling module 1204 is used to identify the groove edge after the wafer laser grooving according to the edge recognition method of any embodiment of the first aspect, and determine whether the groove edge is consistent with the preset wafer laser grooving direction.

[0173] The compensation module 1206 is used to perform precision compensation on the preset wafer laser grooving direction when the groove edge is inconsistent with the preset wafer laser grooving direction.

[0174] Each module in the above-mentioned edge recognition device and wafer laser cutting device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0175] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.13As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an image detection method is implemented.

[0176] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0177] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0178] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0179] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0180] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0181] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An edge recognition method, characterized in that: The method comprises: Acquire a groove image; wherein the groove image is obtained by photographing the groove of the wafer after process processing; Constructing a rotating rectangle on the groove image, and determining a projection direction of the rotating rectangle; the projection direction is parallel to a direction of laser grooving of the wafer; In the groove image, each sampling point in the rotation rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate a projection grayscale curve; The extreme value points on the projected grayscale curve are obtained by taking derivatives according to the projected grayscale curve; the extreme value points are screened, and a search range is determined according to the screened extreme value points; the search range is the area where the groove edge of the wafer is located; the groove edge of the wafer is identified within the search range; when there is only one maximum value point on the projected grayscale curve but no minimum value point, the first-order derivative and the second-order derivative of the projected grayscale curve are obtained to obtain the maximum value point of the first-order derivative of the projected grayscale curve and the minimum value point determined by the second-order derivative, and the minimum value point of the second-order derivative of the projected grayscale curve and the projected The interval of the maximum points of the first-order derivative of the grayscale curve is the search range; when there is a minimum point on the projected grayscale curve, when there is a maximum value, the interval between the maximum point and the minimum point of the projected grayscale curve is the search range; when there is no maximum value, the interval between the minimum point of the projected grayscale curve and the edge of the groove is the search range; when there are multiple minimum points on the projected grayscale curve, the minimum point among the minimum points is obtained according to the projected grayscale curve, and the adjacent maximum point is calculated, and the interval between the minimum point and the maximum point of the projected grayscale curve is the search range.

2. The method according to claim 1, characterized in that The step of constructing a rotating rectangle on the groove image and determining a projection direction of the rotating rectangle; wherein the projection direction is parallel to a direction of laser grooving of the wafer, comprises: Determine a search direction of the rotated rectangle; the search direction is perpendicular to the projection direction and is used to identify the edge of the groove according to the search direction.

3. The method according to claim 1, characterized in that The method further comprises: Arranging a plurality of the rotating rectangles along the laser grooving direction of the wafer; Respectively obtain the projection grayscale curves of the plurality of rotated rectangles to obtain the extreme value points on each of the projection grayscale curves; The groove edge of the wafer is determined according to the connection line between the extreme value points.

4. The method according to claim 2, characterized in that: In the groove image, each sampling point in the rotation rectangle is determined along the projection direction, and the grayscale value of the pixel point corresponding to each sampling point is collected to generate a projection grayscale curve, which includes: Setting a plurality of groups of sampling points along the projection direction in the groove image, wherein each group of sampling points is evenly spaced in the search direction; Sequentially obtain the four pixels closest to each sampling point, and obtain the grayscale value of the sampling point by distance interpolation; The projection grayscale mean value of each group of sampling points in the horizontal projection direction is obtained, and a projection grayscale curve is constructed according to the projection grayscale mean value.

5. The method according to claim 4, characterized in that The step of sequentially acquiring the four pixels closest to each sampling point and obtaining the grayscale value of the sampling point by distance interpolation comprises: Filter the maximum grayscale value and the minimum grayscale value in the rotated rectangle; The grayscale data within the rotated rectangle is normalized according to the maximum grayscale value and the minimum grayscale value.

6. The method according to claim 1, characterized in that The method further comprises: Creating a mask area according to the position of the wafer groove; the mask area is used to cover the area away from the edge of the wafer groove; Perform image grayscale stretching on non-mask areas.

7. A wafer laser cutting method, characterized in that: The method comprises: Laser grooving the wafer according to a preset wafer laser grooving direction; According to the edge recognition method of any one of claims 1 to 6, the edge of the groove after the wafer laser groove is identified, and whether the edge of the groove is consistent with the preset wafer laser groove direction; When the groove edge is inconsistent with the preset wafer laser grooving direction, precision compensation is performed on the preset wafer laser grooving direction.

8. An edge recognition device, characterized in that: The device comprises: An image acquisition module, used to obtain a groove image; wherein the groove image is obtained by photographing the groove of the wafer after process treatment; An initial module is used to construct a rotating rectangle on the groove image and determine a projection direction of the rotating rectangle; the projection direction is parallel to a direction of laser grooving of the wafer; A sampling module, used to determine each sampling point in the rotation rectangle along the projection direction in the groove image, and collect the grayscale value of the pixel point corresponding to each sampling point to generate a projection grayscale curve; The search module is used to determine the extreme point of the projected grayscale curve, and determine the groove edge of the wafer according to the extreme point; obtain the extreme point on the projected grayscale curve by taking the derivative according to the projected grayscale curve; screen the extreme points, and determine the search range according to the screened extreme points; the search range is the area where the groove edge of the wafer is located; identify the groove edge of the wafer within the search range; when there is only one maximum point on the projected grayscale curve but no minimum point, take the first-order derivative and the second-order derivative of the projected grayscale curve to obtain the maximum point of the first-order derivative of the projected grayscale curve and the minimum point determined by the second-order derivative, so as to obtain the minimum point determined by the second-order derivative. The search range is the interval between the minimum point of the second-order derivative of the projected grayscale curve and the maximum point of the first-order derivative of the projected grayscale curve; when there is a minimum point on the projected grayscale curve, when there is a maximum value, the search range is the interval between the maximum point and the minimum point of the projected grayscale curve; when there is no maximum value, the search range is the interval between the minimum point of the projected grayscale curve and the edge of the groove; when there are multiple minimum points on the projected grayscale curve, the minimum point among the minimum points is obtained according to the projected grayscale curve, and the adjacent maximum point is calculated, and the search range is the interval between the minimum point and the maximum point of the projected grayscale curve.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

Images