Wafer double edge detection method, device and equipment

Abstract

The present application relates to a wafer double edge detection method, device and equipment, the method comprising: reading the wafer image to be detected; calculating the image grayscale histogram distribution of the wafer image, and determining the binarization threshold based on the image grayscale histogram distribution; using the determined binarization threshold to binarize the wafer image to obtain a corresponding binarized image; finding each connected area in the binarized image, and screening each connected area, and determining the edge of the wafer image according to the screening result. The purpose of edge detection is achieved by determining the binarization threshold based on the image grayscale histogram distribution during wafer edge detection, and then binarizing the wafer image according to the determined binarization threshold, and then screening the connected area in the binarized image. Compared with the related art, the method of using a fixed area or a variable threshold of the maximum inter-class variance method to binarize the image effectively improves the accuracy of edge detection.

Description

Technical Field

[0001] The present application relates to the technical field of semiconductor device preparation, and in particular to a wafer double-edge detection method, device and equipment. Background Art

[0002] Wafer edge detection has a wide range of applications in the semiconductor manufacturing industry, such as wafer defect detection, wafer positioning, etc. The research on wafer defect detection technology integrates machine vision, digital image processing and other technologies to achieve rapid defect detection on the wafer surface. Wafer edge detection can effectively improve the positioning accuracy of the wafer, thereby ensuring the efficiency and accuracy of chip defect detection. Edge detection is a method that highlights the edge of an image and weakens the image area outside the edge. It can greatly reduce the amount of data processed while retaining useful structural information about the edge of an object, thereby simplifying the image analysis process. The edge of an object is reflected by grayscale discontinuity. In related technologies, commonly used edge detection templates include Sobel operator, Roberts operator, LOG operator, Prewitt operator, Laplacian operator and Canny operator. Among them, when using different edge detection operators to perform edge detection on wafers, a fixed threshold or a varying threshold of the maximum inter-class variance method is used to extract the image edge. Different wafer images have different corresponding edges. When using a fixed threshold or a varying threshold of the maximum inter-class variance method to extract the image edge, the accuracy is usually low, which makes the final edge detection result have a low accuracy. Summary of the invention

[0003] In view of this, the present application proposes a wafer double-edge detection method, which can effectively improve the accuracy of edge detection results.

[0004] According to one aspect of the present application, a wafer double-edge detection method is provided, comprising:

[0005] Read the wafer image to be inspected;

[0006] Calculating an image grayscale histogram distribution of the wafer image, and determining a binarization threshold based on the image grayscale histogram distribution;

[0007] Using the determined binarization threshold to perform binarization processing on the wafer image to obtain a corresponding binarized image;

[0008] Each connected area in the binary image is found, and each connected area is screened, and the edge of the wafer image is determined according to the screening result.

[0009] In a possible implementation manner, determining a binarization threshold based on the image grayscale histogram distribution includes: extracting a first peak value from the image grayscale histogram distribution as the binarization threshold.

[0010] In a possible implementation, when screening the connected regions, the screening is performed according to the area of ​​each connected region.

[0011] In a possible implementation, when the connected regions are screened according to the area of ​​each connected region, the steps include filtering out connected regions whose area is less than or equal to a preset area and retaining connected regions whose area is greater than the preset area.

[0012] In a possible implementation, screening the connected regions further includes: based on the contour concave distance distribution, re-screening the connected regions retained after screening according to the area of ​​the connected regions.

[0013] In a possible implementation, based on the contour concave distance distribution, the connected regions retained after screening according to the area of ​​each of the connected regions are screened again, including:

[0014] Calculate the concave points of the contour line of the connected area;

[0015] Counting the calculated concave points of the contour line of the connected area to obtain the corresponding distribution of the contour concave distance;

[0016] According to the contour concave distance distribution, the best matching connected area is screened out from the connected areas retained after screening according to the area areas of the connected areas, and the contour line of the screened out best matching connected area is determined as the edge of the wafer image.

[0017] In a possible implementation, according to the contour concave distance distribution, the best matching connected region is selected from the connected regions retained after screening according to the area of ​​each connected region, including:

[0018] Comparing the statistically obtained distributions of the contour concave distances, and extracting connected areas where the fluctuation of the contour concave distance distribution is lower than a preset fluctuation;

[0019] From the connected regions where the fluctuation of the contour concave distance distribution is lower than the preset fluctuation, the connected region with the largest area is selected as the most matching connected region.

[0020] In a possible implementation, comparing the statistically obtained contour concave distance distributions and extracting the connected areas where the contour concave distance distribution fluctuation is lower than the preset fluctuation includes:

[0021] Calculate the standard deviation of the profile concave distance distribution;

[0022] The calculated standard deviations are compared with the preset standard deviations, and connected regions having standard deviations lower than the preset standard deviations are extracted according to the comparison results.

[0023] According to another aspect of the present application, there is also provided a wafer double edge detection device, comprising an image reading module, a binarization threshold determination module, a binarization processing module and an edge determination module;

[0024] The image reading module is configured to read the wafer image to be inspected;

[0025] The binarization threshold determination module is configured to calculate the image grayscale histogram distribution of the wafer image and determine the binarization threshold based on the image grayscale histogram distribution;

[0026] The binarization processing module is configured to perform binarization processing on the wafer image using the determined binarization threshold to obtain a corresponding binarized image;

[0027] The edge determination module is configured to find out each connected area in the binary image, screen each connected area, and determine the edge of the wafer image according to the screening result.

[0028] According to another aspect of the present application, a wafer double-edge detection device is also provided, including:

[0029] processor;

[0030] a memory for storing processor-executable instructions;

[0031] Wherein, the processor is configured to implement any of the above-mentioned methods when executing the executable instructions.

[0032] The purpose of edge detection is achieved by determining a binarization threshold based on the image grayscale histogram distribution during wafer edge detection, then binarizing the wafer image according to the determined binarization threshold, and then screening the connected regions in the binarized image. Compared with the related art, which uses a fixed region or a variable threshold of the maximum inter-class variance method to binarize the image, the determination of the binarization threshold is more flexible, and an adaptive binarization threshold is selected for different wafer images for binarization, thereby more effectively improving the accuracy of edge detection.

[0033] Other features and aspects of the present application will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the present application and, together with the description, serve to explain the principles of the present application.

[0035] Figure 1 A flowchart showing a wafer double edge detection method according to an embodiment of the present application is shown;

[0036] Figure 2 Another flow chart showing a wafer double edge detection method according to an embodiment of the present application;

[0037] Figure 3 The figure shows the image grayscale histogram distribution diagram based on which the binarization threshold is selected in the wafer double edge detection method according to one embodiment of the present application;

[0038] Figure 4a and Figure 4b The figures respectively show a wafer image before screening and a wafer image after screening in a process of screening connected regions according to the area of ​​the regions in a wafer double edge detection method according to an embodiment of the present application;

[0039] Figure 5a and Figure 5b The wafer images before and after screening are respectively shown in the process of screening the connected areas according to the contour concave points in the wafer double edge detection method of one embodiment of the present application;

[0040] Figure 6 A structural block diagram of a wafer double edge detection device according to an embodiment of the present application is shown;

[0041] Figure 7 A structural block diagram of a wafer double-edge detection device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0042] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0043] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0044] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0045] Figure 1 FIG. 2 is a flow chart showing a wafer double edge detection method according to an embodiment of the present application. Figure 1 As shown, the method includes: step S100, reading the wafer image to be inspected. Here, it should be noted that the wafer image can be photographed and acquired by using an optical microscope system in an EBI (E-Beam Inspection) device, and then the wafer image can be acquired by reading the EBI device.

[0046] After reading the wafer image, step S200 can be executed to calculate the image grayscale histogram distribution of the wafer image, and determine the binarization threshold based on the image grayscale histogram distribution. That is, after reading the wafer image, when detecting the edge of the wafer, it is necessary to first binarize the wafer image, and then detect the edge of the wafer based on the image after the binarization process. Among them, when performing the binarization process, it is necessary to determine the binarization threshold, so that each pixel in the wafer image is binarized according to the determined binarization threshold. In the method of the embodiment of the present application, the determination of the binarization threshold is carried out by calculating the image grayscale histogram distribution of the wafer image.

[0047] After the binarization threshold is determined, the wafer image can be binarized using the determined binarization threshold in step S300 to obtain a corresponding binarized image. Then, in step S400, each connected region in the binarized image is searched, each connected region is screened, and the edge of the wafer image is determined based on the screening result.

[0048] Therefore, the wafer double edge detection method of the embodiment of the present application determines the binarization threshold based on the image grayscale histogram distribution during wafer edge detection, then binarizes the wafer image according to the determined binarization threshold, and then screens the connected regions in the binarized image to achieve the purpose of edge detection. Compared with the related art, the method of using a fixed area or a variable threshold of the maximum inter-class variance method to binarize the image, the determination of the binarization threshold is more flexible, and an adaptive binarization threshold is selected for different wafer images for binarization, thereby being able to more effectively improve the accuracy of edge detection.

[0049] See also Figure 2 After reading the wafer image to be detected, the wafer image needs to be preprocessed. During the preprocessing, the median filter can be selected to perform image denoising in step S110, and the morphological closing operation can be performed on the wafer image in step S120 to eliminate small holes in the object in the wafer image.

[0050] After the wafer image is preprocessed, the wafer image can be binarized. According to the above, in the wafer double edge detection method of the embodiment of the present application, when the wafer image is binarized, the binarization threshold used is determined based on the image grayscale histogram distribution. Figure 2 and Figure 3 First, through step S210, the image grayscale histogram distribution of the wafer image is calculated, and then, through step S220, the grayscale value after the first peak in the image grayscale histogram distribution is selected as the binarization threshold. Then, through step S300, the wafer image is binarized using the determined binarization threshold to obtain the corresponding binarized image.

[0051] After the wafer image is binarized to obtain a corresponding binary image, step S400 may be executed to find each connected region in the binary image, screen each connected region, and determine the edge of the wafer image according to the screening result.

[0052] Here, it should be noted that in order to effectively improve the accuracy of the screening results, before searching for connected areas in the binary image, it is also necessary to fill the holes in the binary image. Figure 2 , through step S004, the holes in the contour of the binary image are filled. Here, those skilled in the art can understand that the holes in the contour of the image can be filled by conventional technical means in the art, or by a self-designed filling method, which is not specifically limited here.

[0053] In addition, in the wafer double-edge detection method of the embodiment of the present application, when screening connected areas, the connected areas can be screened first according to the area of ​​each connected area to screen out connected areas whose area is larger than a preset area.

[0054] That is, see Figure 2 , Figure 4a and Figure 4b First, through step S410, the area of ​​each connected region is calculated, and then through step S4220, the area of ​​each connected region calculated is compared with the preset area in turn. When the comparison determines that the area of ​​the connected region is less than or equal to the preset area, it indicates that the connected region is an impurity, so through step S430, the connected region is eliminated. When the comparison determines that the area of ​​the connected region is greater than the preset area, it indicates that the contour of the connected region may be the edge of the wafer, so the connected region is retained, and step S440 is executed to extract the contour of the connected region whose area is greater than the preset area.

[0055] Among them, in a possible implementation, the value of the preset area can be flexibly set according to the actual situation. For example, in the wafer double edge detection method of the embodiment of the present application, the value range of the preset area can be: greater than or equal to 0.003*the total area of ​​the wafer image, and less than or equal to 0.008*the total area of ​​the wafer image. Preferably, the value of the preset area can be 0.005*the total area of ​​the wafer image.

[0056] Among them, according to the smooth continuity of the wafer edge, it can be known that the edge of the connected area is a relatively smooth curve. When screening is performed by the area alone, it is not easy to screen out contours with large area and many concave points. Therefore, in the method of the embodiment of the present application, the screening of the connected area also includes the step of fine screening according to the concave points of the contour.

[0057] That is to say, after the above steps, the alternative connected areas (i.e., connected areas whose area is greater than the preset area) are screened out according to the area area of ​​each connected area, in the method of the embodiment of the present application, in order to further improve the accuracy of the wafer inspection results, it also includes a step of re-screening the connected areas (i.e., alternative connected areas) retained after screening according to the area area of ​​each connected area based on the contour recess distance distribution.

[0058] That is, see Figure 2 After extracting the contours of connected regions whose area is larger than the preset area in step S440, the method further includes step S450 of re-screening the connected regions based on the concave distance distribution of the contours of each connected region.

[0059] Among them, see Figure 5a and Figure 5b In step S450, when the connected areas retained after screening according to the area area of ​​each of the connected areas are screened again based on the distribution of contour concave distances, first, it is necessary to perform concave point detection on the contour lines of each connected area and calculate the concave points of the contour lines of the connected areas. Here, it should be pointed out that the method of performing concave point detection on the contour lines of the connected areas and calculating the concave points of the contour lines can be obtained by using convex polygons for detection. The specific process of using convex polygons to perform concave point detection on the contour lines of each connected area can be implemented by conventional technical means in the art, or a corresponding detection algorithm can be designed by itself.

[0060] After the concave points of the contour line of the connected region are calculated, the concave points of the calculated contour line of the connected region are counted using a statistical distribution method to obtain the corresponding contour concave distance distribution. It can be understood by those skilled in the art that the acquisition of the contour concave distance distribution using a statistical distribution method can also be carried out using conventional technical means in the art, so it will not be described in detail here.

[0061] After calculating the contour concave distance distribution of each alternative connected region (i.e., a connected region whose area is larger than a preset area), the best matching connected region can be screened out from the connected regions retained after screening according to the area of ​​each connected region according to the contour concave distance distribution, and the contour line of the screened best matching connected region can be determined as the edge of the wafer image.

[0062] It should be noted that, when selecting the best matching connected region from the connected regions retained after screening according to the area of ​​each connected region according to the contour concave distance distribution, it can be achieved in the following manner.

[0063] That is, first, the statistically obtained distribution of each contour concave distance is compared to extract the connected area whose contour concave distance distribution fluctuation is lower than the preset fluctuation. Then, from the connected areas whose contour concave distance distribution fluctuation is lower than the preset fluctuation, the connected area with the largest area is selected as the most matching connected area.

[0064] Furthermore, when comparing the statistically obtained distributions of the concave distances of each contour and extracting the connected areas where the fluctuation of the distribution of the concave distance of the contour is lower than the preset fluctuation, this can be achieved by means of standard deviation.

[0065] That is, see Figure 2 , by calculating the standard deviation of the distribution of the concave distances of each contour, and then in step S460, comparing the calculated standard deviations with the preset standard deviations, and extracting the connected areas with standard deviations lower than the preset standard deviations according to the comparison results. Then, in step S470, the connected areas with the largest area are screened out from the connected areas with standard deviations lower than the preset standard deviations, and the contour line of the connected area is used as the wafer edge.

[0066] It should be noted that the calculation of the standard deviation of the concave distance distribution of each contour can be achieved by using a conventional standard deviation calculation formula, which will not be explained here with examples.

[0067] Therefore, the wafer double edge detection method of the embodiment of the present application analyzes the grayscale distribution curve of the image for different edge images, selects adaptive thresholds for different images for binarization, and combines the shape features of the connected domain to formulate a wafer double edge extraction strategy, which effectively improves the accuracy of target edge selection. Compared with other edge detection algorithms, the wafer double edge detection method of the embodiment of the present application can flexibly select the binary edge threshold according to the grayscale distribution of the image, and secondary screen out the wafer edge based on the characteristics of the image area and concave point distribution, which effectively eliminates the interference of noise and has strong robustness.

[0068] It should be noted that although Figures 1 to 5bThe above-mentioned wafer double edge detection method is introduced as an example, but those skilled in the art will understand that the present application should not be limited thereto. In fact, the user can flexibly set the specific implementation method of each step according to personal preferences and / or actual application scenarios, as long as different binarization thresholds can be flexibly selected to binarize the wafer image, and the connected areas can be screened according to the area of ​​the connected areas and the distribution of the contour concave distance.

[0069] Accordingly, based on any of the above, the present application also provides a wafer double edge detection device. Since the working principle of the wafer double edge detection device provided by the present application is the same or similar to the principle of the wafer double edge detection method provided by the present application, the repeated parts will not be repeated.

[0070] See also Figure 6 The wafer double edge detection device 100 provided in the present application includes an image reading module 110, a binarization threshold determination module 120, a binarization processing module 130 and an edge determination module 140. Among them, the image reading module 110 is configured to read the wafer image currently to be detected. The binarization threshold determination module 120 is configured to calculate the image grayscale histogram distribution of the wafer image, and determine the binarization threshold based on the image grayscale histogram distribution. The binarization processing module 130 is configured to use a determined binarization threshold to perform binarization processing on the wafer image to obtain a corresponding binarized image. The edge determination module 140 is configured to find out each connected area in the binary image, and screen each connected area, and determine the edge of the wafer image according to the screening result.

[0071] Furthermore, according to another aspect of the present application, a wafer double edge detection device 200 is also provided. Figure 7 The wafer double edge detection device 200 of the embodiment of the present application includes a processor 210 and a memory 220 for storing executable instructions of the processor 210. The processor 210 is configured to implement any of the above-mentioned wafer double edge detection methods when executing the executable instructions.

[0072] Here, it should be noted that the number of processors 210 may be one or more. At the same time, in the wafer double edge detection device 200 of the embodiment of the present application, an input device 230 and an output device 240 may also be included. Among them, the processor 210, the memory 220, the input device 230 and the output device 240 may be connected through a bus or in other ways, which are not specifically limited here.

[0073] The memory 220 is a computer-readable storage medium that can be used to store software programs, computer executable programs, and various modules, such as the program or module corresponding to the wafer double edge detection method of the embodiment of the present application. The processor 210 executes various functional applications and data processing of the wafer double edge detection device 200 by running the software program or module stored in the memory 220.

[0074] The input device 230 may be used to receive input numbers or signals. The signals may be key signals related to user settings and function control of the device / terminal / server. The output device 240 may include display devices such as display screens.

[0075] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A wafer double edge detection method, characterized in that: include: Read the wafer image to be inspected; Calculating an image grayscale histogram distribution of the wafer image, and determining a binarization threshold based on the image grayscale histogram distribution; Using the determined binarization threshold to perform binarization processing on the wafer image to obtain a corresponding binarized image; Finding each connected region in the binary image, and screening the connected regions according to the area of ​​each connected region, and re-screening the connected regions retained after screening according to the area of ​​each connected region based on the contour concave distance distribution to determine the edge of the wafer image; The method of re-screening the connected regions retained after screening according to the area of ​​each connected region based on the contour concave distance distribution to determine the edge of the wafer image includes: Calculate the concave points of the contour line of the connected area; Counting the calculated concave points of the contour line of the connected area to obtain the corresponding distribution of the contour concave distance; According to the contour concave distance distribution, the best matching connected area is screened out from the connected areas retained after screening according to the area areas of the connected areas, and the contour line of the screened out best matching connected area is determined as the edge of the wafer image.

2. The method according to claim 1, characterized in that Determining a binarization threshold based on the image grayscale histogram distribution includes: extracting a first peak value from the image grayscale histogram distribution as the binarization threshold.

3. The method according to claim 1, characterized in that When the connected regions are screened according to the area of ​​each connected region, the steps include filtering out the connected regions whose area is less than or equal to the preset area and retaining the connected regions whose area is greater than the preset area.

4. The method according to claim 1, characterized in that: According to the contour concave distance distribution, the best matching connected area is selected from the connected areas retained after screening according to the area areas of the connected areas, including: Comparing the statistically obtained distributions of the contour concave distances, and extracting connected areas where the fluctuation of the contour concave distance distribution is lower than a preset fluctuation; From the connected regions where the fluctuation of the contour concave distance distribution is lower than the preset fluctuation, the connected region with the largest area is selected as the most matching connected region.

5. The method according to claim 4, characterized in that The statistically obtained contour concave distance distributions are compared to extract the connected areas where the contour concave distance distribution fluctuation is lower than the preset fluctuation, which includes: Calculate the standard deviation of the profile concave distance distribution; The calculated standard deviations are compared with the preset standard deviations, and connected regions having standard deviations lower than the preset standard deviations are extracted according to the comparison results.

6. A wafer double edge detection device, characterized in that: It includes an image reading module, a binarization threshold determination module, a binarization processing module and an edge determination module; The image reading module is configured to read the wafer image to be inspected; The binarization threshold determination module is configured to calculate the image grayscale histogram distribution of the wafer image and determine the binarization threshold based on the image grayscale histogram distribution; The binarization processing module is configured to perform binarization processing on the wafer image using the determined binarization threshold to obtain a corresponding binarized image; The edge determination module is configured to find out each connected area in the binary image, and screen the connected areas according to the area area of ​​each connected area, calculate the concave points of the contour line of the connected area, count the concave points of the calculated contour line of the connected area, and obtain the corresponding contour concave distance distribution. According to the contour concave distance distribution, the best matching connected area is screened out from the connected areas retained after screening according to the area area of ​​each connected area, and the contour line of the screened best matching connected area is determined to be the edge of the wafer image.

7. A wafer double edge detection device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the method described in any one of claims 1 to 5 when executing the executable instructions.

Images