Wafer appearance detection method and system, computer and storage medium

Abstract

The invention provides a wafer appearance detection method and system, a computer and a storage medium, and the method comprises the following steps: obtaining image data from a cloud, screening coordinate data corresponding to a data abnormal value of each crystal grain from a first detection result, and obtaining the detection data of each peripheral crystal grain of adjacent crystal grains; if the number of the marked crystal grains to be detected is greater than a first preset value, judging that the adjacent crystal grains are abnormal crystal grains; and obtaining a plurality of second detection results comprising a plurality of abnormal crystal grains. On the basis of traditional detection equipment, abnormal crystal grains in image data are marked, whether the adjacent crystal grains are abnormal or not is judged based on the distribution condition of peripheral crystal grains of the adjacent crystal grains corresponding to the adjacent region of the abnormal crystal grains, and according to the region continuous distribution characteristic caused by artificial damage, the abnormal crystal grains are detected. The problem that a traditional optical detection scheme cannot detect slight damage can be solved, and the detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of wafer detection, and particularly to a wafer appearance detection method, system, computer and storage medium. Background Art

[0002] Wafer appearance defect detection refers to detecting the wafer surface to discover and identify defects therein to ensure the quality and stability of the wafer. The existence of appearance defects will affect the crystal growth, electrical performance of the wafer and the reliability of the final product.

[0003] In the prior art, mainly by taking pictures on the machine tool and transmitting the images to the optical system device for detection through image acquisition; because the scratches and contaminations caused by humans have a gradual change in degree caused by the contact force, the abnormal degrees in the middle and severe can be detected normally, but the slight defects cannot be detected by conventional traditional detection, resulting in a low detection accuracy. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the purpose of the present invention is to provide a wafer appearance detection method, system, computer and storage medium, aiming to solve the technical problem of low detection accuracy in the prior art.

[0005] To achieve the above purpose, in the first aspect, the present invention provides: A wafer appearance detection method, including the following steps: Obtain the image data of the wafer according to the CCD camera, upload the image data to the cloud, and output a first detection result based on the optical detection device; Obtain the image data from the cloud, screen the coordinate data corresponding to the data outliers of each grain from the first detection result, and perform a marking process on the marked grains corresponding to the coordinate data in the image data; Obtain the detection data of each peripheral grain of the adjacent grains corresponding to the adjacent area of the marked grains based on the first detection result, and mark the peripheral grains with abnormal detection data as the to-be-tested marked grains; If the number of the to-be-tested marked grains is greater than a first preset value, determine that the adjacent grains are abnormal grains; Repeat the above steps of obtaining the detection data of each peripheral grain of the adjacent grains corresponding to the adjacent area of the abnormal grains and determining whether the number of the to-be-tested marked grains is greater than the first preset value until the number of the to-be-tested marked grains is less than the first preset value, and finally obtain a second detection result including several abnormal grains, and output a target test result based on the second test result.

[0006] According to one aspect of the above technical solution, after the step of performing a marking process on the marked grains corresponding to the coordinate data in the image data, the method further includes: Calculating the parabola slope of a line segment formed by a plurality of continuous marked grains in the image data, and extending the line segment based on the parabola slope to obtain an extended line segment, wherein the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of the adjacent marked grain is less than a second preset value; The grains in the corresponding area on the extended line segment are all determined to be abnormal grains.

[0007] According to one aspect of the above technical solution, after the step of obtaining the image data from the cloud, the method further includes: A detection model is established based on a deep learning algorithm, and various electrode defect images are selected to train the model until the accuracy of the detection model reaches the target requirement; The image data is input into the detection model to identify and obtain electrode abnormality data of each grain.

[0008] According to one aspect of the above technical solution, the grain includes an N-electrode region, a P-electrode region, a cutting region, a light-emitting region, a Finger region and a Mesa region, and the data anomaly value includes the electrode abnormal data, the cutting region abnormal data, the light-emitting region abnormal data, the Finger region abnormal data and the Mesa region abnormal data.

[0009] According to one aspect of the above technical solution, the step of outputting a target test result based on the second test result specifically includes: If the first test result corresponding to the target die is normal and the second test result is abnormal, outputting the target test result according to the second test result; If the second test result corresponding to the target die and the second test result are both abnormal, the target test result is output according to the first test result.

[0010] In the second aspect, the present solution also provides a wafer appearance inspection system, including: A first detection module, configured to obtain image data of the wafer using a CCD camera, upload the image data to the cloud, and output a first detection result based on an optical detection device; a marking module, configured to obtain the image data from the cloud, select coordinate data corresponding to abnormal data values ​​of each grain from the first detection result, and mark the marked grains corresponding to the coordinate data in the image data; An expansion module, based on the first detection result, obtains detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the marked grain, and marks the peripheral grain with abnormal detection data as the marked grain to be tested; A second detection module, if the number of the marked grains to be detected is greater than a first preset value, determines that the adjacent grains are abnormal grains; An output module is used to repeat the above steps of obtaining the detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the abnormal grain, and judging whether the number of the marked grains to be tested is greater than the first preset value, until the number of the marked grains to be tested is less than the first preset value, and finally obtaining a plurality of second detection results including a plurality of abnormal grains, and outputting a target test result based on the second test result.

[0011] According to one aspect of the above technical solution, the system further includes: A connection module, configured to calculate a parabola slope of a line segment formed by a plurality of continuous marked grains in the image data, and to extend the line segment based on the parabola slope to obtain an extended line segment, wherein the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of the adjacent marked grain is less than a second preset value; The grains in the corresponding area on the extended line segment are all determined to be abnormal grains.

[0012] According to one aspect of the above technical solution, the system further includes: The detection module is used to establish a detection model based on a deep learning algorithm and select various electrode defect images to train the model until the accuracy of the detection model meets the target requirements; The image data is input into the detection model to identify and obtain electrode abnormality data of each grain.

[0013] In a third aspect, the present invention provides a computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the wafer appearance inspection method as described in the above technical solution when executing the computer program.

[0014] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, the wafer appearance inspection method as described in the above technical solution is implemented.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: by marking the abnormal grains in the image data on the basis of traditional detection equipment, the distribution of the peripheral grains of the adjacent grains corresponding to the adjacent areas of the abnormal grains is used to determine whether the adjacent grains are abnormal, and the grains around the abnormal grains are secondary judged according to the continuous distribution characteristics of the area caused by human damage. This can solve the problem that traditional optical detection schemes cannot detect minor damage and improve the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Flow chart of the wafer appearance inspection method in the first embodiment of the present invention; Figure 2 Schematic diagram of the distribution of line segments in the image data in the first embodiment of the present invention; Figure 3 Electrode defect image of the crystal grains in the first embodiment of the present invention.

[0017] Figure 4 Block diagram of the structure of the wafer appearance detection system in the second embodiment of the present invention; Figure 5 Schematic diagram of the hardware structure of the computer in the third embodiment of the present invention; The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. Specific Embodiments

[0018] For ease of understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided so that the disclosure of the present invention is thorough and comprehensive.

[0019] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be an intermediate element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intermediate element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0021] Please refer to Figure 1 , which shows the flowchart of a wafer appearance detection method in the first embodiment of the present invention. As Figure 1 shown, the method includes the following steps: Step S100: Obtain the image data of the wafer according to the CCD camera, upload the image data to the cloud, and output a first detection result based on the optical detection device.

[0022] Step S200: Obtain the image data from the cloud, screen the coordinate data corresponding to the data outliers of each crystal grain from the first detection result, and perform a marking process on the marked crystal grains corresponding to the coordinate data in the image data.

[0023] Specifically, in some application scenarios of this embodiment, the above-mentioned crystal grains include an N electrode region, a P electrode region, a cutting region, a light-emitting region, a Finger region, and a Mesa region, and the data outliers include electrode anomaly data, cutting region anomaly data, light-emitting region anomaly data, Finger region anomaly data, and Mesa region anomaly data. When the data of the corresponding crystal grain in the image data is normal, the display color is green. Different data outliers represent different types of anomalies in the crystal grain, and the corresponding marker display colors in the image data are also different.

[0024] Step S300: Obtain the detection data of each peripheral crystal grain of the adjacent crystal grains corresponding to the marked crystal grain based on the first detection result, and mark the peripheral crystal grains with abnormal detection data as the to-be-tested marked crystal grains.

[0025] Step S400: If the number of the to-be-tested marked crystal grains is greater than a first preset value, determine that the adjacent crystal grains are abnormal crystal grains.

[0026] Specifically, in some application scenarios of this embodiment, the above-mentioned peripheral crystal grains refer to the crystal grains in the adjacent region of the marked crystal grain, that is, the eight crystal grains outside the marked crystal grain; the first preset value is preferably 4, that is, when the first test result of the adjacent crystal grains in the adjacent region of the marked crystal grain is normal, but there are 4 or more crystal grains with abnormal detection data outside it, then determine that the peripheral crystal grain is an abnormal crystal grain.

[0027] Step S500: Repeat the steps of obtaining the detection data of each peripheral crystal grain of the adjacent crystal grains corresponding to the abnormal crystal grain and determining whether the number of the to-be-tested marked crystal grains is greater than the first preset value until the number of the to-be-tested marked crystal grains is less than the first preset value, and finally obtain a second detection result including several abnormal crystal grains, and output a target test result based on the second test result.

[0028] Preferably, in this embodiment, after the step of marking the marked crystal grain corresponding to the coordinate data in the image data for the loss of the scratch type, the method further includes: Calculate the parabola slope of the line segment formed by several consecutive marked crystal grains in the image data, extend the line segment based on the parabola slope to obtain an extended line segment, and the distance between the coordinates of the marked crystal grain corresponding to the end point of the extended line segment and the coordinates of its adjacent marked crystal grain is less than a second preset value; Determine that all the crystal grains in the corresponding region of the extended line segment are abnormal crystal grains.

[0029] For easy understanding, there are many types of the above-mentioned data outliers, and different types of outliers have different color manifestations in the image data. The above-mentioned electrode abnormal data includes N-electrode abnormal data and P-electrode abnormal data. Taking the P-electrode abnormality as an example, an image area of red (R:G:B = 255, 0, 0) is selected, the distribution of the images in this area is marked, the line segments with "continuous line shapes (straight lines & curves, etc.)" are marked, and the parabola slope is obtained by taking the coordinates of three points among them.

[0030] As Figure 2 shown, the line segment corresponding to the P-electrode abnormal data passes through the coordinates of points A, B, and C. The coefficients a, b, and c can be obtained through y = ax² + bx + c, and then the grain coordinates that the parabolic line of the unified abnormal return value will pass through can be simulated. Under the condition that the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of its adjacent marked grain is less than the second preset value, the coordinates corresponding to the coordinates of the extended line segment are determined as NG to improve the determination accuracy and reduce the abnormal outflow from the client.

[0031] Specifically, in this embodiment, after the step of obtaining the image data from the cloud for the needle mark detection of the electrode, the method further includes: Establishing a detection model based on a deep learning algorithm, selecting various types of electrode defect images to train the model until the accuracy of the detection model reaches the target requirement; Inputting the image data into the detection model to identify the electrode abnormal data of each grain.

[0032] Specifically, in this embodiment, after obtaining the image data from the cloud, AI detection can be used to quickly process the images, and cooperate with traditional optical detection equipment for synchronous detection to improve the detection efficiency; The principle of the needle mark detection model includes the following steps: Marking the defects on all the electrode CB rings as must-pass marks, then guiding the training through a large amount of data to improve the accuracy of the needle mark negative samples, and then performing reverse filtering deep learning by matching the negative samples to achieve the automatic detection of electrode defects.

[0033] As Figure 3 shown, in this embodiment, various defect images of the abnormal classes and difference classes of the grain electrodes are presented.

[0034] Preferably, in this embodiment, the step of outputting the target test result based on the second test result specifically includes: If the first test result corresponding to the target grain is normal and the second test result is abnormal, then output the target test result according to the second test result; If the second test result corresponding to the target die and the second test result are both abnormal, the target test result is output according to the first test result.

[0035] Furthermore, in the present embodiment, the above-mentioned first test result and the second test result both include normal, or display one of the above-mentioned data abnormality values. When the analysis of the first test result and the second test result are both abnormal, and the types of the data abnormality values ​​are different, the output result of the traditional equipment is mainly used; for example, the first test result of the target grain shows that there is abnormal contamination of the electrode, and the second test result shows that there is slight contamination in the light-emitting area. In order to display the problem more comprehensively, the first test result is maintained as the target test result for output in the present embodiment; it can be understood that in other embodiments of the present application, it can also be displayed by setting the priority of the data abnormality, that is, outputting the target test result.

[0036] In summary, the wafer appearance inspection method in the above-mentioned embodiment of the present invention, by marking the abnormal grains in the image data on the basis of traditional inspection equipment, determines whether the adjacent grains are abnormal based on the distribution of the peripheral grains of the adjacent grains corresponding to the adjacent areas of the abnormal grains, and performs a secondary judgment on the grains around the abnormal grains according to the continuous distribution characteristics of the area caused by human damage. This can solve the problem that traditional optical inspection schemes cannot detect minor damages and improve the detection accuracy.

[0037] like Figure 4 As shown, the second embodiment of the present invention provides a wafer appearance inspection system, including a first inspection module 100, a marking module 200, an expansion module 300, a second inspection module 400 and an output module 500; The first detection module 100 is used to obtain image data of the wafer according to the CCD camera, upload the image data to the cloud, and output a first detection result based on the optical detection equipment; The marking module 200 is used to obtain the image data from the cloud, select the coordinate data corresponding to the abnormal data value of each grain from the first detection result, and mark the marked grains corresponding to the coordinate data in the image data; The expansion module 300 acquires the detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the marked grain based on the first detection result, and marks the peripheral grain with abnormal detection data as the marked grain to be tested; The second detection module 400 determines that the adjacent grains are abnormal grains if the number of the marked grains to be detected is greater than a first preset value; The output module 500 is configured to repeat the above steps of obtaining the detection data of each peripheral grain of the adjacent grains corresponding to the adjacent regions of the abnormal grains, and determining whether the number of the to-be-tested marked grains is greater than a first preset value until the number of the to-be-tested marked grains is less than the first preset value, and finally obtaining a second detection result including a plurality of abnormal grains, and outputting a target test result based on the second test result.

[0038] Preferably, in this embodiment, the system further includes: A connection module, configured to calculate the parabola slope of a line segment formed by a plurality of consecutive marked grains in the image data, extend the line segment based on the parabola slope to obtain an extended line segment, and the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of the adjacent marked grain is less than a second preset value; Determine that the grains in the corresponding region on the extended line segment are all abnormal grains.

[0039] Preferably, in this embodiment, the system further includes: A detection module, configured to establish a detection model based on a deep learning algorithm, select various types of electrode defect images to train the model until the accuracy of the detection model reaches the target requirement; Input the image data into the detection model to identify the electrode abnormal data of each grain.

[0040] It should be noted that each of the above modules can be a functional module or a program module, and can be implemented either by software or by hardware. For the modules implemented by hardware, each of the modules can be located in the same processor; or each of the modules can also be located in different processors in any combined form.

[0041] Embodiment III The third embodiment of the present application provides a computer, which may include a processor 81 and a memory 82 storing computer program instructions.

[0042] Specifically, the above-mentioned processor 81 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0043] Among them, the memory 82 may include a mass memory for data or commands. By way of example and not limitation, the memory 82 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 82 may include removable or non-removable (or fixed) media. Where appropriate, the memory 82 may be internal or external to the data processing device. In a particular embodiment, the memory 82 is non-volatile memory. In a particular embodiment, the memory 82 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable read-only memory (EAROM), or a flash memory (FLASH), or a combination of two or more of these. Where appropriate, the RAM may be a static random access memory (SRAM) or a dynamic random access memory (DRAM), where the DRAM may be a fast page mode dynamic random access memory (FPMDRAM), an extended date out dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0044] The memory 82 can be used to store or cache various data files required for processing and / or communication, as well as possible computer program commands executed by the processor 81.

[0045] The processor 81 reads and executes the computer program commands stored in the memory 82 to implement any one of the wafer appearance detection methods in the above embodiments.

[0046] In some of the embodiments, the computer may further include a communication interface 83 and a bus 80. Among them, as Figure 5 shown, the processor 81, the memory 82, and the communication interface 83 are connected through the bus 80 and complete communication with each other.

[0047] The communication interface 83 is used to implement communication between the various modules, devices, units, and / or devices in the embodiments of the present application. The communication interface 83 can also implement data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations, etc.

[0048] Bus 80 includes hardware, software, or both, and couples components of a computer to each other. Bus 80 includes, but is not limited to, at least one of the following: Data Bus, Address Bus, Control Bus, Expansion Bus, Local Bus. By way of example and not limitation, Bus 80 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable bus or a combination of two or more of these. In a suitable case, Bus 80 may include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0049] Embodiment 4 The fourth embodiment of the present application provides a readable storage medium. Computer program instructions are stored on the readable storage medium; when the computer program instructions are executed by a processor, any one of the wafer appearance detection methods in the above embodiments is implemented.

[0050] The technical features of the above-described embodiments may be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0051] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A wafer appearance inspection method, characterized in that: The following steps are involved: Acquire image data of the wafer according to the CCD camera, upload the image data to the cloud, and output a first detection result based on the optical detection device; Acquire the image data from the cloud, select coordinate data corresponding to abnormal data values ​​of each grain from the first detection result, and mark the marked grains corresponding to the coordinate data in the image data; Based on the first detection result, the detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the marked grain are acquired, and the peripheral grains with abnormal detection data are marked as marked grains to be tested; If the number of the marked grains to be tested is greater than a first preset value, the adjacent grains are determined to be abnormal grains; Repeat the above steps of obtaining the detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the abnormal grain, and judging whether the number of the marked grains to be tested is greater than the first preset value, until the number of the marked grains to be tested is less than the first preset value, and finally obtain a plurality of second detection results including a plurality of abnormal grains, and output the target test results based on the second test results.

2. The wafer appearance inspection method according to claim 1, characterized in that: After the step of marking the marked grains corresponding to the coordinate data in the image data, the method further comprises: Calculating the parabola slope of a line segment formed by a plurality of continuous marked grains in the image data, and extending the line segment based on the parabola slope to obtain an extended line segment, wherein the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of the adjacent marked grain is less than a second preset value; The grains in the corresponding area on the extended line segment are all determined to be abnormal grains.

3. The wafer appearance inspection method according to claim 1, characterized in that: After the step of acquiring the image data from the cloud, the method further includes: A detection model is established based on a deep learning algorithm, and various electrode defect images are selected to train the model until the accuracy of the detection model reaches the target requirement; The image data is input into the detection model to identify and obtain electrode abnormality data of each grain.

4. The wafer appearance inspection method according to claim 3, characterized in that: The crystal grain includes an N-electrode region, a P-electrode region, a cutting region, a light-emitting region, a Finger region and a Mesa region, and the data abnormal value includes the electrode abnormal data, the cutting region abnormal data, the light-emitting region abnormal data, the Finger region abnormal data and the Mesa region abnormal data.

5. The wafer appearance inspection method according to claim 4, characterized in that: The step of outputting a target test result based on the second test result specifically includes: If the first test result corresponding to the target die is normal and the second test result is abnormal, outputting the target test result according to the second test result; If the second test result corresponding to the target die and the second test result are both abnormal, the target test result is output according to the first test result.

6. A wafer appearance inspection system, characterized in that: include: A first detection module, configured to obtain image data of the wafer using a CCD camera, upload the image data to the cloud, and output a first detection result based on an optical detection device; a marking module, configured to obtain the image data from the cloud, select coordinate data corresponding to abnormal data values ​​of each grain from the first detection result, and mark the marked grains corresponding to the coordinate data in the image data; An expansion module, based on the first detection result, obtains detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the marked grain, and marks the peripheral grain with abnormal detection data as the marked grain to be tested; A second detection module, if the number of the marked grains to be detected is greater than a first preset value, determines that the adjacent grains are abnormal grains; An output module is used to repeat the above steps of obtaining the detection data of each peripheral grain of the adjacent grain corresponding to the adjacent area of ​​the abnormal grain, and judging whether the number of the marked grains to be tested is greater than the first preset value, until the number of the marked grains to be tested is less than the first preset value, and finally obtaining a plurality of second detection results including a plurality of abnormal grains, and outputting a target test result based on the second test result.

7. The wafer appearance inspection system according to claim 6, characterized in that: The system further comprises: A connection module, configured to calculate a parabola slope of a line segment formed by a plurality of continuous marked grains in the image data, and to extend the line segment based on the parabola slope to obtain an extended line segment, wherein the distance between the coordinates of the marked grain corresponding to the end point of the extended line segment and the coordinates of the adjacent marked grain is less than a second preset value; The grains in the corresponding area on the extended line segment are all determined to be abnormal grains.

8. The wafer appearance inspection system according to claim 6, characterized in that: The system further comprises: The detection module is used to establish a detection model based on a deep learning algorithm and select various electrode defect images to train the model until the accuracy of the detection model meets the target requirements; The image data is input into the detection model to identify and obtain electrode abnormality data of each grain.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the wafer appearance inspection method according to any one of claims 1 to 5 is implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the wafer appearance inspection method as described in any one of claims 1 to 5 is implemented.

Images