Wafer anomaly detection method and device, computer device and storage medium

Abstract

The application relates to a wafer anomaly detection method and device, computer equipment and a storage medium, wherein the method comprises the following steps: obtaining a scanning image of a wafer to be detected from a detection device; performing region segmentation processing on the scanning image to obtain a plurality of region segmentation images; determining the position information of a first die in the region segmentation image corresponding to the first die, and performing feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain target features of the first die; comparing the target features of the first die with reference features in a preset data set to determine whether the first die is abnormal or not; and determining a detection result of the wafer to be detected according to the dies with abnormalities in the wafer to be detected, wherein the detection result is used to indicate the abnormal type of the wafer to be detected. According to the application, the position of the abnormal die in the wafer can be accurately detected, and the precision of wafer anomaly detection is improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductors, and more particularly to a wafer anomaly detection method, apparatus, computer device, and computer-readable storage medium. Background Technology

[0002] In semiconductor manufacturing, wafer anomaly detection is a crucial step. Its purpose is to detect potential defects within the wafer, enabling timely monitoring of product quality during production. Automated optical inspection (AOI) equipment acquires scanned images of the wafer and identifies potential defects within these images. However, traditional AOI equipment has limited detection capabilities for specific anomalies. For example, for products with varying grayscale values, traditional AOI equipment only performs grayscale analysis, which can easily lead to over-detection and difficulty in accurately locating defects, resulting in low detection accuracy. Therefore, improving the accuracy of wafer anomaly detection is a pressing issue that needs to be addressed. Summary of the Invention

[0003] In view of the shortcomings of the prior art, the purpose of this application is to provide a wafer anomaly detection method, apparatus, computer equipment and storage medium, which can accurately detect the location of abnormal grains in the wafer, thereby improving the accuracy of wafer anomaly detection.

[0004] A wafer anomaly detection method includes: acquiring a scanned image of a wafer to be inspected from an inspection device; performing region segmentation processing on the scanned image to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one die of the wafer to be inspected; determining the position information of a first die in the region segmentation image corresponding to the first die, and performing feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain target features of the first die; the first die being any die in the wafer to be inspected; comparing the target features of the first die with reference features in a preset dataset to determine whether the first die has an anomaly or not; and determining the detection result of the wafer to be inspected based on the presence of an abnormal die in the wafer to be inspected, the detection result being used to indicate the anomaly type of the wafer to be inspected.

[0005] The aforementioned wafer anomaly detection method divides the scanned image of the wafer under inspection into multiple segmented images. Then, feature extraction is performed on the grains within each segmented image to identify anomalous grains within the wafer. Based on the presence of these anomalous grains, the detection result of the wafer is determined. Compared to traditional AOI equipment that only performs grayscale judgment, this method offers greater robustness and can more accurately locate defects within the wafer, thus improving the accuracy of wafer anomaly detection.

[0006] Optionally, the scanned image of the wafer to be inspected includes N partial images of the wafer, where N is a positive integer greater than 1. Obtaining the scanned image of the wafer from the inspection device includes: when it is determined that the temporary storage path of the inspection device is empty, sequentially obtaining the partial images of the wafer from the inspection device; and when it is determined that the target storage path of the inspection device contains result files corresponding to the N partial images, confirming that N partial images have been obtained from the inspection device. Therefore, in this technical solution, by sequentially obtaining the partial images of the wafer to be inspected from the inspection device, the scanned image of the wafer can be obtained in a more flexible manner, which is beneficial to enhancing the flexibility of wafer anomaly detection.

[0007] Optionally, the scanned image is subjected to region segmentation processing to obtain multiple region-segmented images, including: performing region segmentation processing on the scanned image based on a segmentation threshold to obtain multiple region-segmented images. It is evident that in this technical solution, using a segmentation threshold to segment the scanned image helps improve the efficiency of the segmentation process.

[0008] Optionally, determining the position information of the first chip in the region segmentation image corresponding to the first chip includes: obtaining the context information of each pixel in the region segmentation image corresponding to the first chip; and determining the position information of the first chip in the region segmentation image corresponding to the first chip based on the context information of each pixel in the region segmentation image corresponding to the first chip. It is evident that in this technical solution, combining the context information of each pixel in the region segmentation image corresponding to the first chip helps to more accurately determine the position information of the first chip in the region segmentation image corresponding to the first chip.

[0009] Optionally, the target features of the first wafer are compared with reference features in a preset dataset to determine whether the first wafer has an anomaly or not. This includes: calculating the feature similarity between the target features of the first wafer and the reference features in the preset dataset; and determining whether the first wafer has an anomaly or not based on the feature similarity between the target features of the first wafer and the reference features in the preset dataset. Therefore, in this technical solution, determining whether the first wafer has an anomaly based on the feature similarity between the target features of the first wafer and the reference features in the preset dataset helps improve the intelligence of wafer anomaly detection.

[0010] Optionally, based on the presence of abnormal grains in the wafer under inspection, the inspection result of the wafer under inspection is determined, including: determining the location information of the abnormal grains in the wafer under inspection; and determining the anomaly type of the wafer under inspection based on the matching degree between the location information of the abnormal grains in the wafer under inspection and the location information corresponding to each anomaly type. It can be seen that in this technical solution, matching the location information of the abnormal grains in the wafer under inspection with the location information corresponding to each anomaly type can make wafer anomaly detection more precise.

[0011] Optionally, the above wafer anomaly detection method further includes: outputting the location information of the abnormal grains in the wafer under test, as well as the detection result of the wafer under test. It is evident that in this technical solution, outputting the detection result of the wafer under test provides a more intuitive display of the wafer anomaly detection result.

[0012] Based on the same inventive concept, this application also provides a wafer anomaly detection device, which includes: an acquisition module for acquiring a scanned image of a wafer to be inspected from an inspection device; a processing module for performing region segmentation processing on the scanned image to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one die of the wafer to be inspected; determining the position information of a first die in the region segmentation image corresponding to the first die, and performing feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain the target feature of the first die; the first die is any die in the wafer to be inspected; comparing the target feature of the first die with the reference feature in a preset dataset to determine whether the first die has an anomaly or not; and determining the detection result of the wafer to be inspected based on the dies with an anomaly in the wafer to be inspected, the detection result being used to indicate the anomaly type of the wafer to be inspected.

[0013] Based on the same inventive concept, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described wafer anomaly detection method.

[0014] Based on the same inventive concept, embodiments of this application provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of a computer device reads and executes the computer instructions from the computer-readable storage medium, causing the computer device to perform the above-described wafer anomaly detection method. Attached Figure Description

[0015] Figure 1A This is a schematic diagram of the system architecture applied to embodiments of this application;

[0016] Figure 1B This application provides a schematic diagram of the hardware and software architecture of a wafer inspection system.

[0017] Figure 2 A schematic flowchart of a wafer anomaly detection method provided in an embodiment of this application;

[0018] Figure 3 A schematic flowchart illustrating the process of acquiring a partial image of a wafer to be inspected from an inspection device, as provided in an embodiment of this application;

[0019] Figure 4 A schematic flowchart of another wafer anomaly detection method provided in an embodiment of this application;

[0020] Figure 5 This is a schematic diagram of the structure of a wafer anomaly detection device provided in an embodiment of this application;

[0021] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application.

[0022] Explanation of reference numerals in the attached figures:

[0023] 102-Computer equipment; 103-Wafer; 104-Inspection equipment; 401-Scanned image of the wafer to be inspected; 41-Local image; 50-Wafer anomaly detection device; 501-Acquisition module; 502-Processing module; 60-Computer equipment; 601-Communication interface; 602-Memory; 603-Processor. Detailed Implementation

[0024] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0026] In semiconductor manufacturing, traditional AOI (Automated Optical Inspection) equipment acquires scanned images of the wafer and identifies the locations of potential defects, enabling anomaly detection. However, traditional AOI equipment has limited detection capabilities for certain specific anomalies. For example, for products with grayscale variations, traditional AOI equipment only performs grayscale analysis on the wafer's scanned image, failing to accurately detect defect locations and resulting in low detection accuracy.

[0027] Therefore, this application aims to provide a solution that can solve the above-mentioned technical problems, the details of which will be described in subsequent embodiments.

[0028] Please see Figure 1A , Figure 1A This is a schematic diagram of the wafer anomaly detection system architecture applied in an embodiment of this application. Figure 1A As shown, the system includes computer equipment 102 and detection equipment 104. It should be understood that... Figure 1A The number and form of the devices shown are for illustrative purposes only and do not constitute a limitation on the embodiments of this application.

[0029] Computer device 102 can be a terminal or a server. A terminal can be a tablet, laptop, desktop computer, etc. A server can be a standalone server, a server cluster consisting of several servers, or a cloud computing service center. This application does not limit the scope of the application.

[0030] The inspection device 104 can acquire scanned images of the wafer 103. The inspection device 104 can communicate with the computer device 102, either via wired or wireless communication. The inspection device 104 can send the acquired scanned images of the wafer 103 to the computer device 102, enabling the computer device 102 to obtain the scanned images of the wafer 103.

[0031] Please see Figure 1B , Figure 1B This is a schematic diagram of the hardware and software architecture of a wafer inspection system provided in an embodiment of this application. Figure 1B As shown, the hardware components may include testing equipment, a network interface card (NIC), and a computer. The testing equipment, configured with a NIC, can communicate with the computer. The software components may include data transmission software, image processing software, wafer anomaly detection software, and test result output software.

[0032] The data transmission software can send scanned images of the wafer under inspection acquired by the inspection equipment to a computer without affecting the normal operation of the inspection equipment. The image processing software can perform region segmentation processing on the scanned images of the wafer under inspection, obtaining multiple segmented images. The wafer anomaly detection software can locate each die in the multiple segmented images of the wafer under inspection, and perform feature extraction processing on each die based on its location information to obtain the target features of each die; then, it compares the target features of each die with reference features in a preset dataset to determine the presence of abnormal dies in the wafer under inspection. The detection result output software can determine the detection result of the wafer under inspection based on the presence of abnormal dies, and output the location information of the abnormal dies in the wafer under inspection, as well as the detection result of the entire wafer.

[0033] The wafer anomaly detection method provided in the embodiments of this application is described below.

[0034] Please see Figure 2 , Figure 2 This is a schematic flowchart illustrating a wafer anomaly detection method provided in an embodiment of this application. This method can be applied to... Figure 1A In the system architecture shown, execution is performed by computer device 102, such as... Figure 2 As shown, the method includes, but is not limited to, the following steps:

[0035] S201, Obtain a scanned image of the wafer to be inspected from the inspection equipment.

[0036] In this embodiment, the computer device can establish a wired or wireless communication connection with the testing device. After establishing the communication connection, the computer device can acquire a scanned image of the wafer to be tested from the testing device.

[0037] The inspection equipment can be an AOI (Optical Image Inspection) device. AOI devices can detect wafer defects based on optical principles. When inspecting a wafer, an AOI device first scans the wafer with a camera to obtain a scanned image, and then inspects the scanned image to determine potential defects. However, AOI devices can only identify a limited number of defect types and have low detection accuracy. This application embodiment utilizes an AOI device to acquire a scanned image of the wafer to be inspected, which is then inspected by a computer to obtain more accurate inspection results. It should be noted that this application embodiment uses an AOI device as an example; it is understood that the inspection device can also be other devices with scanning capabilities, such as optical character recognition (OCR) devices.

[0038] In some embodiments, the inspection device can automatically send the scanned image of the wafer to be inspected to a computer device, enabling the computer device to acquire the scanned image of the wafer. Alternatively, the inspection device can respond to a request message from the computer device by sending the scanned image of the wafer to be inspected to the computer device.

[0039] In some embodiments, the scanned image of the wafer to be inspected may include N partial images of the wafer, where N is a positive integer greater than 1. That is, the inspection device can scan the wafer by region to obtain multiple partial images of the wafer. Here, combining the multiple partial images of the wafer obtained from the scan can yield the scanned image of the entire wafer.

[0040] In a practical implementation, the inspection device can store the scanned partial images of the wafer to be inspected in a temporary storage path. Upon completion of scanning one wafer, the temporary storage path is cleared, and then the scanning of the next wafer begins, storing the resulting partial image of the next wafer in the same temporary storage path. Thus, the computer device can sequentially acquire partial images of the wafer to be inspected from the inspection device when it determines that the temporary storage path of the inspection device is cleared. For example, the inspection device can send a first notification message to the computer device when the temporary storage path is cleared, informing the computer device that it can begin acquiring partial images of the wafer to be inspected. Here, the order in which the computer device acquires the partial images of the wafer to be inspected from the inspection device can be based on the scanning order of N partial images. For example, if the inspection device first scans and obtains the first partial image of the wafer to be inspected, and then scans and obtains the second partial image, the computer device will first acquire the first partial image of the wafer to be inspected from the inspection device, and then acquire the second partial image.

[0041] It should be understood that the first local image and the second local image are two of N local images, and no specific limitation is made here.

[0042] It is worth noting that the inspection equipment can inspect partial images of the wafer to be inspected and store the reference inspection results corresponding to the partial images of the wafer to be inspected in the target storage path of the inspection equipment. Therefore, when the computer equipment determines that the target storage path of the inspection equipment contains result files corresponding to N partial images of the wafer to be inspected, it can determine that N partial images of the wafer to be inspected have been obtained from the inspection equipment, that is, it can determine that the acquisition of the scanned images of the wafer to be inspected is complete, and end the acquisition of the scanned images of the wafer to be inspected.

[0043] For example, please see Figure 3 , Figure 3 This is a schematic flowchart illustrating a process for acquiring a partial image of a wafer to be inspected from an inspection device, as provided in an embodiment of this application. Figure 3 As shown, since the inspection equipment clears the temporary storage path when it starts scanning a wafer, the computer equipment can determine that the temporary storage path of the inspection equipment is cleared as a start signal to begin acquiring partial images of the wafer to be inspected. Then, the computer equipment can sequentially acquire partial images of the wafer to be inspected from the temporary storage path of the inspection equipment according to the scanning order of the inspection equipment. The acquisition of partial images of the wafer to be inspected is stopped when the computer equipment determines that the target storage path of the inspection equipment contains result files corresponding to N partial images of the wafer to be inspected. In other words, the acquisition of partial images of the wafer to be inspected is considered to have ended when the target storage path of the inspection equipment generates result files corresponding to N partial images of the wafer to be inspected. That is, the generation of N partial images of the wafer to be inspected in the target storage path of the inspection equipment indicates that the inspection equipment has completed scanning the wafer to be inspected.

[0044] S202, Perform region segmentation processing on the scanned image to obtain multiple region segmented images.

[0045] In this embodiment of the application, after the computer device acquires a scanned image of the wafer to be inspected, it can perform region segmentation processing on the scanned image to obtain multiple region segmentation images. One of the multiple region segmentation images is used to display at least one grain of the wafer to be inspected. For example, the first region segmentation image can be used to display the first and second grains of the wafer to be inspected, and the second region segmentation image can be used to display the third grain of the wafer to be inspected.

[0046] It should be understood that the first region segmentation image and the second region segmentation image are two of a total of multiple region segmentation images, and no specific limitation is made here.

[0047] In some embodiments, the computer device performs region segmentation processing on the scanned image by segmenting the scanned image based on a segmentation threshold to obtain multiple segmented region images. The segmentation threshold can be a dynamic threshold, and the segmentation threshold may differ for scanned images of different wafers. Optionally, during the region segmentation process, pixels in the scanned image with grayscale values ​​greater than or equal to the segmentation threshold can be represented by a first grayscale value, such as 255. Pixels in the scanned image with grayscale values ​​less than the segmentation threshold can be represented by a second grayscale value, such as 0. This method can segment the scanned image of the wafer to be inspected into several meaningful region segmentation images, facilitating subsequent feature extraction based on these region segmentation images.

[0048] In some embodiments, when the illumination distribution in the scanned image is uniform, the segmentation threshold used for region segmentation processing of the scanned image can be a global threshold. This global threshold can be applied to the entire scanned image; that is, a fixed segmentation threshold can be used for region segmentation processing of the entire scanned image. Optionally, the segmentation threshold used for region segmentation processing of the scanned image can also be a local threshold. The local threshold can be calculated for each pixel of the scanned image, or it can be calculated for each local image of the scanned image. Using such a local threshold, region segmentation processing of the scanned image can be performed more effectively when the illumination distribution in the scanned image is uneven (e.g., there are shadows or the light has a sloping gradient distribution). Optionally, the segmentation threshold used for region segmentation processing of the scanned image can be calculated using a machine learning model.

[0049] S203, determine the position information of the first grain in the region segmentation image corresponding to the first grain, and perform feature extraction processing on the first grain based on the position information of the first grain in the region segmentation image corresponding to the first grain to obtain the target features of the first grain.

[0050] In this embodiment, after obtaining multiple region segmentation images, the computer device can determine the position information of the first die in the region segmentation image corresponding to the first die, and perform feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain the target features of the first die. The position information of the first die in the region segmentation image corresponding to the first die can be the coordinate information of the first die in the region segmentation image corresponding to the first die. Here, the first die can be any die in the wafer to be inspected.

[0051] In some embodiments, the computer device determines the position information of the first chip in the region segmentation image corresponding to the first chip by first acquiring the context information of each pixel in the region segmentation image corresponding to the first chip, and then determining the position information of the first chip in the region segmentation image corresponding to the first chip based on the context information of each pixel in the region segmentation image corresponding to the first chip. Here, the context information can be the interaction information between pixels. For example, if there are pixels 1, 2, and 3 in the region segmentation image corresponding to the first chip, then the context information between pixels 1 and 2, between pixels 2 and 3, and between pixels 1 and 3 can be calculated. Optionally, the computer device can acquire the context information of each pixel in the region segmentation image corresponding to the first chip through a trained neural network.

[0052] In some embodiments, the computer device performs feature extraction processing on the first chip based on the position information of the first chip in the segmented image of the region corresponding to the first chip. This can be achieved by learning a feature representation of the first chip based on its position information in the segmented image of the region corresponding to the first chip; that is, the feature representation of the first chip can be used as the target feature of the first chip. The feature representation of the first chip can be a feature vector corresponding to the first chip, a set of feature vectors corresponding to the first chip, or a feature vector map corresponding to the first chip; this embodiment does not limit the specific features. Optionally, when learning the feature representation of the first chip, the computer device can also combine the context information of each pixel in the segmented image of the region corresponding to the first chip. For example, the position information of the first chip in the segmented image of the region corresponding to the first chip, and the context information of each pixel in the segmented image of the region corresponding to the first chip, can be input into a trained neural network for feature extraction processing to obtain the target feature of the first chip. Furthermore, the computer device can continuously collect the target features of the chip to train the neural network, enabling the neural network to acquire the depth features of the chip, resulting in a more efficient and accurate feature representation of the chip.

[0053] S204, compare the target features of the first grain with the reference features in the preset dataset to determine whether the first grain has an anomaly or not.

[0054] In this embodiment, after obtaining the target features of the first grain, the computer device compares the target features of the first grain with reference features in a preset dataset to determine whether the first grain has an anomaly or not. The preset dataset can be obtained from a database and can be a single dataset or multiple datasets. The reference features in the preset dataset can be, for example, features of grains with anomalies or features of grains without anomalies.

[0055] In some embodiments, the computer device compares the target features of a first die with reference features in a preset dataset to determine whether the first die exhibits an anomaly or not. This may include: calculating the feature similarity between the target features of the first die and the reference features in the preset dataset; and determining whether the first die exhibits an anomaly or not based on the feature similarity between the target features of the first die and the reference features in the preset dataset. The feature similarity between the target features of the first die and the reference features in the preset dataset indicates the degree of similarity between the target features of the first die and the reference features in the preset dataset. A higher feature similarity between the target features of the first die and the reference features in the preset dataset indicates a higher degree of similarity between the target features of the first die and the reference features in the preset dataset.

[0056] In some embodiments, when the reference features in the preset dataset are features of aberrant grains, if the feature similarity between the target features of the first grain and the reference features in the preset dataset is greater than or equal to a first threshold, it indicates that the target features of the first grain are highly similar to the features of aberrant grains, and thus the first grain can be determined to be aberrant. Conversely, if the feature similarity between the target features of the first grain and the reference features in the preset dataset is less than the first threshold, it can be determined that the first grain is not aberrant. Similarly, when the reference features in the preset dataset are features of aberrant grains, if the feature similarity between the target features of the first grain and the reference features in the preset dataset is greater than or equal to a second threshold, it indicates that the target features of the first grain are highly similar to the features of aberrant grains, and thus the first grain is not aberrant. Conversely, if the feature similarity between the target features of the first grain and the reference features in the preset dataset is less than the second threshold, it can be determined that the first grain is aberrant.

[0057] The first threshold and the second threshold may be the same or different. The first threshold and the second threshold may be preset fixed thresholds or dynamically changing thresholds; this application does not limit this.

[0058] S205, Based on the presence of abnormal grains in the wafer to be tested, determine the test result of the wafer to be tested.

[0059] In this embodiment of the application, if the computer device determines that the wafer to be inspected contains abnormal grains, it can determine the inspection result of the wafer based on the presence of abnormal grains. The inspection result of the wafer to be inspected can be used to indicate the type of abnormality. The type of abnormality of the wafer to be inspected may include, but is not limited to, foreign objects, scratches, defects (such as protrusions, misalignments, or missing parts), contamination, and defective raw materials.

[0060] In some embodiments, the computer device can determine the anomaly type of the wafer under test based on the matching degree between the location information of the abnormal dies in the wafer and the location information corresponding to each anomaly type. This is because the anomaly type of the wafer under test is related to the location of the abnormal dies in the wafer. By matching the location information of the abnormal dies in the wafer with the location information corresponding to each anomaly type, if the location information of the abnormal dies in the wafer successfully matches the location information corresponding to the target anomaly type, the anomaly type of the wafer under test can be determined to be the target anomaly type. Here, a successful match between the location information of the abnormal dies in the wafer and the location information corresponding to the target anomaly type can mean that the location information of the abnormal dies in the wafer is the same as the location information corresponding to the target anomaly type, or it can mean that the location information of the abnormal dies in the wafer is similar to the location information corresponding to the target anomaly type. Optionally, the computer device can also directly input the location information of the abnormal dies in the wafer to be inspected into the artificial intelligence (AI) model, and the AI ​​model can output the abnormality type of the wafer to be inspected.

[0061] In some embodiments, the computer device may also output the location information of the abnormal die in the wafer under test, as well as the test result of the wafer under test. Optionally, the computer device may provide voice prompts regarding the location information of the abnormal die in the wafer under test, as well as the test result of the wafer under test. For example, it may output the voice prompt message "There is an abnormal die in the upper left corner of the wafer; the wafer may be scratched." The computer device may also display the location information of the abnormal die in the wafer under test, as well as the test result of the wafer under test. For example, it may highlight the location information of the abnormal die in the wafer under test and display the test result of the wafer under test in the form of a pop-up window.

[0062] In some embodiments, the wafer defect detection method of this application can be divided into two stages. Please refer to [link to relevant documentation]. Figure 4 , Figure 4 This is a schematic flowchart illustrating another wafer anomaly detection method provided in an embodiment of this application. Figure 4As shown, the computer device can acquire a scanned image 401 of the wafer to be inspected, which may include N local images 41. For each local image 41, processing can be performed in two stages: ① Grain localization: The computer device can acquire a segmentation threshold for the local image 41, and perform region segmentation processing on the local image 41 based on the segmentation threshold to obtain multiple region segmentation images; then, the context information of each pixel in the region segmentation image corresponding to the first grain is acquired, so that the position information of the first grain in the region segmentation image corresponding to the first grain can be determined based on the context information of each pixel in the region segmentation image corresponding to the first grain, thus achieving grain localization. The first grain is any grain in the wafer to be inspected. ② Defect judgment: Based on the position information of the first grain in the region segmentation image corresponding to the first grain, feature extraction processing is performed on the first grain to obtain the target features of the first grain; then, the target features of the first grain are compared with the reference features in a preset dataset to determine whether the first grain has an anomaly or not. Finally, after processing all N local images using the above two stages, the detection result of the wafer to be inspected can be determined based on the presence of abnormal grains in the wafer. This detection result can be used to indicate the type of abnormality of the wafer to be inspected.

[0063] By implementing the embodiments of this application, the scanned image of the wafer to be inspected can be divided into multiple segmented images. Then, feature extraction processing is performed on the grains in each segmented image to identify abnormal grains within the wafer. Based on the presence of abnormal grains, the inspection result of the wafer is determined. This method obtains the depth features of the grains in the wafer, which, compared to traditional AOI equipment that only performs grayscale judgment, has stronger robustness and can more accurately locate the position of defects in the wafer, thus improving the accuracy of wafer anomaly detection.

[0064] It is understandable that since inspection equipment (such as traditional AOI equipment) can inspect the wafer to be inspected and obtain a reference inspection result for the wafer, this reference inspection result can be used to indicate whether the wafer to be inspected has an anomaly or not. For wafers to be inspected that have an anomaly, the embodiments of this application can be used to further refine the inspection of the wafer with the anomaly to determine the type of anomaly. In other words, traditional AOI equipment can classify whether the wafer to be inspected has an anomaly. For wafers to be inspected that are classified as having an anomaly, their scanned images can be sent to a computer. By executing the method provided in the embodiments of this application, a computer device can divide a scanned image of a wafer to be inspected that is classified as abnormal into multiple region segmentation images. Then, feature extraction processing is performed on the grains in each region segmentation image to determine the presence of abnormal grains in the wafer to be inspected that is classified as abnormal. Finally, based on the presence of abnormal grains in the wafer to be inspected, the specific abnormality type (such as foreign matter, scratches, defects, etc.) of the wafer to be inspected that is classified as abnormal is determined, thereby achieving refined wafer abnormality detection and improving the reliability and accuracy of wafer abnormality detection.

[0065] Please see Figure 5 This is a schematic diagram of the structure of a wafer anomaly detection device provided in an embodiment of this application. The wafer anomaly detection device 50 includes:

[0066] The acquisition module 501 is used to acquire scanned images of the wafer to be inspected from the inspection equipment;

[0067] The processing module 502 is used to perform region segmentation processing on the scanned image to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one grain of the wafer to be inspected; the position information of the first grain in the region segmentation image corresponding to the first grain is determined, and feature extraction processing is performed on the first grain based on the position information of the first grain in the region segmentation image corresponding to the first grain to obtain the target features of the first grain; the first grain is any grain in the wafer to be inspected; the target features of the first grain are compared with the reference features in the preset dataset to determine whether the first grain has an anomaly or not; based on the grains in the wafer to be inspected that have an anomaly, the detection result of the wafer to be inspected is determined, and the detection result is used to indicate the anomaly type of the wafer to be inspected.

[0068] Optionally, the scanned image of the wafer to be inspected includes N partial images of the wafer to be inspected, where N is a positive integer greater than 1; the processing module 502 is further configured to, when it is determined that the temporary storage path of the inspection device is in an empty state, sequentially obtain the partial images of the wafer to be inspected from the inspection device; and when it is determined that the target storage path of the inspection device contains result files corresponding to N partial images, determine that N partial images have been obtained from the inspection device.

[0069] Optionally, the processing module 502 is also used to perform region segmentation processing on the scanned image based on the segmentation threshold of the scanned image to obtain multiple region segmented images.

[0070] Optionally, the acquisition module 501 is further configured to acquire the context information of each pixel in the region segmentation image corresponding to the first grain; the processing module 502 is further configured to determine the position information of the first grain in the region segmentation image corresponding to the first grain based on the context information of each pixel in the region segmentation image corresponding to the first grain.

[0071] Optionally, the processing module 502 is further configured to calculate the feature similarity between the target features of the first grain and the reference features in the preset dataset; and determine whether the first grain has an anomaly or not based on the feature similarity between the target features of the first grain and the reference features in the preset dataset.

[0072] Optionally, the processing module 502 is further configured to determine the location information of the abnormal dies in the wafer to be inspected; and to determine the abnormal type of the wafer to be inspected based on the matching degree between the location information of the abnormal dies in the wafer to be inspected and the location information corresponding to each abnormality type.

[0073] Optionally, the processing module 502 is also used to output the location information of the abnormal grains in the wafer to be tested, as well as the detection result of the wafer to be tested.

[0074] It should be noted that, Figure 5 For details not mentioned in the corresponding embodiments and the specific implementation methods of each step, please refer to [link to relevant documentation]. Figures 2-4 The embodiments shown and the foregoing content will not be repeated here.

[0075] Please see Figure 6This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. The computer device 60 may include a communication interface 601, a memory 602, and a processor 603. The communication interface 601, memory 602, and processor 603 are connected via one or more communication buses, which are used to enable communication between these components. The communication interface 601 may include a standard wired interface or a wireless interface (such as a WIFI interface). The memory 602 may include volatile memory, such as random-access memory (RAM); the memory 602 may also include non-volatile memory, such as flash memory, solid-state drive (SSD), etc.; the memory 602 may also include a combination of the above types of memory. The processor 603 may be a central processing unit (CPU). The processor 603 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), etc. The aforementioned PLD can be a field-programmable gate array (FPGA), a generic array logic (GAL), etc.

[0076] Optionally, the memory 602 is also used to store program instructions, which the processor 603 can also call to implement the wafer anomaly detection method in the embodiments of this application.

[0077] Optionally, the processor 603 calls program instructions stored in the memory 602 to: acquire a scanned image of the wafer to be inspected from the inspection device; perform region segmentation processing on the scanned image to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one die of the wafer to be inspected; determine the position information of the first die in the region segmentation image corresponding to the first die, and perform feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain the target feature of the first die; the first die is any die in the wafer to be inspected; compare the target feature of the first die with the reference feature in the preset dataset to determine whether the first die has an anomaly or not; determine the inspection result of the wafer to be inspected based on the die with an anomaly in the wafer to be inspected, and the inspection result is used to indicate the anomaly type of the wafer to be inspected.

[0078] Optionally, the processor 603 may also call the program instructions to: when it is determined that the temporary storage path of the inspection device is in an empty state, sequentially obtain partial images of the wafer to be inspected from the inspection device; when it is determined that there are result files corresponding to N partial images in the target storage path of the inspection device, determine that N partial images have been obtained from the inspection device.

[0079] Optionally, the processor 603 may also call the program instructions to perform region segmentation processing on the scanned image based on the segmentation threshold of the scanned image, thereby obtaining multiple region segmented images.

[0080] Optionally, the processor 603 may also call the program instructions to: obtain the context information of each pixel in the region segmentation image corresponding to the first die; and determine the position information of the first die in the region segmentation image corresponding to the first die based on the context information of each pixel in the region segmentation image corresponding to the first die.

[0081] Optionally, the processor 603 may also call the program instructions to: calculate the feature similarity between the target features of the first grain and the reference features in the preset dataset; and determine whether the first grain has an anomaly or not based on the feature similarity between the target features of the first grain and the reference features in the preset dataset.

[0082] Optionally, the processor 603 may also call the program instructions to: determine the location information of the abnormal dies in the wafer to be inspected; and determine the abnormal type of the wafer to be inspected based on the matching degree between the location information of the abnormal dies in the wafer to be inspected and the location information corresponding to each abnormal type.

[0083] Optionally, the processor 603 may also call the program instructions to: output the location information of the abnormal dies in the wafer to be tested, and the detection result of the wafer to be tested.

[0084] It should be understood that the principles and beneficial effects of the computer device 60 in solving the problem described in the embodiments of this application are the same as those in this application. Figures 2-4 The embodiments shown and the foregoing content have similar principles and beneficial effects in solving the problem, and will not be repeated here for the sake of brevity.

[0085] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method described in the embodiments of this application.

[0086] Accordingly, embodiments of this application provide a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. The processor of a computer device 60 reads and executes the computer instructions from the computer-readable storage medium, causing the computer device 60 to perform the method described in the embodiments of this application.

[0087] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0088] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0089] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A wafer anomaly detection method, characterized in that, The method includes: Obtain scanned images of the wafer to be inspected from the inspection equipment; The scanned image is subjected to region segmentation processing to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one grain of the wafer to be inspected; The position information of the first grain in the region segmentation image corresponding to the first grain is determined, and feature extraction processing is performed on the first grain based on the position information of the first grain in the region segmentation image corresponding to the first grain to obtain the target features of the first grain; the first grain is any grain in the wafer to be detected; The target features of the first grain are compared with the reference features in the preset dataset to determine whether the first grain has an anomaly or not. Based on the presence of abnormal grains in the wafer to be inspected, the inspection result of the wafer to be inspected is determined, and the inspection result is used to indicate the type of abnormality of the wafer to be inspected.

2. The wafer anomaly detection method as described in claim 1, characterized in that, The scanned image of the wafer to be inspected includes N partial images of the wafer to be inspected, where N is a positive integer greater than 1; The step of acquiring a scanned image of the wafer to be inspected from the inspection equipment includes: When it is determined that the temporary storage path of the testing equipment is in an empty state, partial images of the wafer to be tested are sequentially obtained from the testing equipment; When it is determined that the target storage path of the detection device contains the result files corresponding to the N local images, it is determined that the N local images have been obtained from the detection device.

3. The wafer anomaly detection method as described in claim 1, characterized in that, The process of performing region segmentation on the scanned image yields multiple region-segmented images, including: Based on the segmentation threshold of the scanned image, the scanned image is processed to segment the region, resulting in multiple region-segmented images.

4. The wafer anomaly detection method as described in claim 3, characterized in that, Determining the position information of the first grain in the region segmentation image corresponding to the first grain includes: Obtain the context information of each pixel in the region segmentation image corresponding to the first grain; Based on the context information of each pixel in the region segmentation image corresponding to the first chip, the position information of the first chip in the region segmentation image corresponding to the first chip is determined.

5. The wafer anomaly detection method as described in claim 1, characterized in that, The first step involves comparing the target features of the first grain with the reference features in a preset dataset. Determining whether the first grain has an anomaly or not includes: Calculate the feature similarity between the target features of the first grain and the reference features in the preset dataset; Based on the feature similarity between the target features of the first grain and the reference features in the preset dataset, it is determined whether the first grain has an anomaly or not.

6. The wafer anomaly detection method according to any one of claims 1-5, characterized in that, The step of determining the detection result of the wafer under test based on the presence of abnormal grains in the wafer under test includes: Determine the location information of the abnormal grains in the wafer to be inspected; The abnormality type of the wafer to be inspected is determined based on the matching degree between the location information of the abnormal grains in the wafer to be inspected and the location information corresponding to each abnormality type.

7. The wafer anomaly detection method as described in claim 6, characterized in that, The method further includes: Output the location information of the abnormal grains in the wafer to be tested, as well as the detection result of the wafer to be tested.

8. A wafer anomaly detection device, characterized in that, The wafer anomaly detection device includes: The acquisition module is used to acquire scanned images of the wafer to be inspected from the inspection equipment; The processing module is used to perform region segmentation processing on the scanned image to obtain multiple region segmentation images; wherein, one of the multiple region segmentation images is used to display at least one die of the wafer to be inspected; determine the position information of the first die in the region segmentation image corresponding to the first die, and perform feature extraction processing on the first die based on the position information of the first die in the region segmentation image corresponding to the first die to obtain the target features of the first die; the first die is any die of the wafer to be inspected; compare the target features of the first die with the reference features in the preset dataset to determine whether the first die has an anomaly or not; determine the detection result of the wafer to be inspected based on the die with an anomaly in the wafer to be inspected, and the detection result is used to indicate the anomaly type of the wafer to be inspected.

9. A computer device, characterized in that, include: The processor, the communication interface, and the memory are interconnected, wherein the memory stores executable program code, and the processor is used to call the executable program code to implement the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1-7.

Images