Real-time flaw detection method and equipment applying AI algorithm, medium and product

Abstract

The invention provides a real-time flaw detection method and device applying an AI algorithm, a medium and a product. The method comprises the following steps: after a wafer is tested by using a probe, acquiring a test area image; identifying the test area image by using an AI algorithm to obtain a probe mark detection result in the test area image; evaluating the probe mark detection result by adopting a preset detection standard, and determining whether the probe mark has defects or not; and under the condition that the probe mark has the defect, determining a processing decision according to the defect type. According to the technical scheme, whether the probe has defects or not can be detected on the basis of the probe marks, the probe defects can be rapidly and accurately detected through the image recognition technology, the probe detection efficiency is improved, the process is automatic, and more manpower resources do not need to be input.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a real-time flaw detection method, equipment, medium and product using an AI algorithm. Background Art

[0002] In recent years, with the rapid development of science and technology, the wafer production industry, as the basic unit of the semiconductor industry, has developed rapidly.

[0003] During the chip probe test before wafer testing and packaging, corresponding electrical, logic and functional tests must be performed to detect whether they meet product specifications. This process requires the probe of the probe card to contact the pad on the wafer before an electrical connection can be established with the chip in the wafer, thereby achieving the purpose of electrical performance testing. If the probe does not work properly or too much pressure is applied, it will cause damage to the chip. In addition, it will also cause damage to the probe, resulting in huge economic losses and affecting detection efficiency. The traditional method of detecting probes is often based on manual detection. Technicians will regularly use microscopes to manually check the probe cards, which is laborious and time-consuming. Therefore, how to scientifically and efficiently detect probes is a technical problem that needs to be solved urgently. Summary of the invention

[0004] One purpose of the present application is to provide a real-time flaw detection method, equipment, medium and product using an AI algorithm, at least to solve the problems of low efficiency and high human resource consumption of traditional probe detection methods. The purpose of the present application is to provide a new real-time flaw detection method using an AI algorithm. After using a probe to test a wafer, the method obtains a test area image and uses an AI algorithm to identify the test area image to obtain a probe mark detection result in the test area image, and uses a preset detection standard to evaluate the probe mark detection result to determine whether the probe mark has a defect, so as to determine a processing decision based on the defect type when the probe mark has a defect. By adopting this solution, it is possible to detect whether the probe has a defect based on the probe mark, so as to quickly and accurately detect probe defects through image recognition technology, thereby improving the detection efficiency of the probe, and automating the process without investing more human resources.

[0005] To achieve the above objectives, some embodiments of the present application provide the following aspects:

[0006] In a first aspect, some embodiments of the present application further provide a real-time flaw detection method using an AI algorithm, the method comprising:

[0007] After testing the wafer using the probe, acquiring a test area image;

[0008] Using an AI algorithm to identify the test area image, and obtain a probe marking detection result in the test area image;

[0009] Using a preset detection standard to evaluate the probe mark detection result to determine whether the probe mark has a defect;

[0010] In the event that the probe marks the presence of a defect, a processing decision is determined based on the defect type.

[0011] In a second aspect, some embodiments of the present application further provide an electronic device, comprising: one or more processors; and a memory storing computer program instructions, wherein the computer program instructions, when executed, cause the processor to perform the steps of the method described above.

[0012] In a third aspect, some embodiments of the present application further provide a computer-readable medium having computer program instructions stored thereon, wherein the computer program instructions can be executed by a processor to implement the method as described above.

[0013] In a fourth aspect, some embodiments of the present application further provide a computer program product, comprising a computer program / instruction, which implements the steps of the method described above when executed by a processor.

[0014] Compared with the prior art, in the solution provided in the embodiment of the present application, after the wafer is tested with a probe, the test area image is acquired and the test area image is identified using an AI algorithm to obtain the probe mark detection result in the test area image, and the probe mark detection result is evaluated using a preset detection standard to determine whether the probe mark has a defect, so as to determine the processing decision according to the defect type when the probe mark has a defect. By adopting this solution, it is possible to detect whether the probe has a defect based on the probe mark, so as to quickly and accurately detect probe defects through image recognition technology, thereby improving the detection efficiency of the probe, and automating the process without investing more human resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flowchart of a real-time flaw detection method using an AI algorithm provided in Example 1 of the present application;

[0016] Figure 2 It is a flowchart of a real-time flaw detection method using an AI algorithm provided in Example 2 of the present application;

[0017] Figure 3 An exemplary structural diagram of the electronic device is disclosed. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] The technical solution provided by the embodiment of the present application is described in detail below through specific embodiments and their application scenarios in conjunction with the accompanying drawings.

[0020] Embodiment 1

[0021] Figure 1 : is a flow chart of a real-time flaw detection method using an AI algorithm provided in Example 1 of the present application. Figure 1 As shown, the specific steps include:

[0022] Step S101, after testing the wafer using a probe, acquiring a test area image;

[0023] The wafer is tested by using a probe, which can be connected to the pad on the wafer to achieve electrical connection with the chip on the wafer, and then the connectivity and logic tests can be performed. Specifically, it can be used to screen qualified chips, identify qualified chips before the wafer is sent to the packaging factory, avoid the subsequent packaging process of bad chips, and save costs. It can also be used for characteristic evaluation and parameter monitoring, and the electrical parameters of the device or circuit are evaluated. Engineers maintain the quality level of the process by monitoring the distribution state of the parameters, such as monitoring the threshold voltage, current gain and other parameters of the transistor to ensure that they are within the design specification range. In addition, calculation and feedback can be performed through testing. The calculation of qualified and defective chips can provide comprehensive performance feedback to wafer production personnel, such as statistical yield, defective product type and distribution information, to help production personnel understand the production status and make improvements.

[0024] The test area image can be acquired through the camera. The test area can be the entire wafer, one or more chips on the wafer, or a region including only one pad. It can be understood that the range of the test area image here can be related to the distance between the camera and the wafer, and to the shooting field of view of the camera.

[0025] In this embodiment, optionally, a probe is used to test the wafer, including:

[0026] Control the probe to contact the pad surface;

[0027] Controlling movement of a preset distance when the probe contacts the pad surface;

[0028] Probe marking is formed by contact and movement.

[0029] In the process of controlling the probe to move toward the wafer, real-time detection is performed. When it is detected that the probe contacts the pad surface, the probe stops moving toward the wafer and controls the probe to move laterally on the pad surface, so as to scratch the oxide layer on the pad surface and form a probe mark. The probe mark may be a scratch mark formed by the lateral movement of the probe on the pad surface. It is understandable that by moving laterally, a better connection between the probe and the pad can be achieved, thereby avoiding the influence of the oxide layer on the subsequent test process and thus affecting the accuracy of the test results.

[0030] Specifically, the following steps may be included: each wafer contains many chips, and each chip surface is attached with a pad; with the help of a probe card, the probe is lowered onto the wafer until the probe tip contacts the wafer pad; the test signal may be transmitted between the test instrument and the chip; the test is performed according to the test items, and then the good chips may be separated from the bad chips.

[0031] The movement of the probe as it contacts the pad surface is also important. The surface of each pad is covered with a very thin layer of oxide material. In order to make contact with the metal underneath, the probe tip must break through this thin oxide. Because the probe is very thin, it bends when it contacts the wafer. As the probe bends, the probe tip slides or scrubs across the pad surface. This scrubbing creates a probe mark. This scrubbing action causes the probe tip to break through the surface oxide, helping to form good electrical contact with the metal underneath.

[0032] Through such a setting, the present solution can not only perform accurate testing, but also leave probe marks on the surface of the pad for detecting abnormalities of the probe.

[0033] Step S102, using an AI algorithm to identify the test area image to obtain a probe mark detection result in the test area image;

[0034] The recognition process of the test area image may be performed using an AI algorithm to determine whether there is a probe mark in the test area image, as well as the position, size, and depth of the probe mark.

[0035] The AI ​​algorithm for image recognition of the test area image may include: Convolutional Neural Network (CNN), which automatically extracts features from the image through the convolution layer, uses different convolution kernels to slide on the image for local perception, and captures local features of the image, such as edges, textures, etc.; the pooling layer is used to reduce the dimension of the feature map, reduce the amount of calculation and the risk of overfitting; the fully connected layer integrates the extracted features to achieve classification or recognition tasks. Convolutional neural networks have certain invariance to the translation, rotation, and scaling of images, can effectively extract local and global features of images, and perform well in tasks such as image classification, target detection, and image segmentation. Convolutional neural networks can be widely used in various image recognition fields, such as face recognition, object recognition, and medical image analysis. For example, they are used to identify specific people or abnormal behaviors in security monitoring; and to identify roads, vehicles, pedestrians, etc. in autonomous driving. Recurrent Neural Network (RNN), whose structure includes neurons with recurrent connections, can retain previous information when processing sequence data, and process time-related data such as image sequences or video frame sequences by modeling information on time series. Recurrent neural networks are good at processing sequence data and have advantages in processing information with time dimension, such as image sequences. Recurrent neural networks are used in image description generation, video content understanding, etc. For example, a description of video content is generated based on a series of image frames. Generative Adversarial Network (GAN) consists of a generator and a discriminator. The generator tries to generate realistic images to deceive the discriminator, while the discriminator strives to distinguish between real images and images generated by the generator. The two continuously improve their respective capabilities through adversarial training. Generative adversarial networks can generate new image data with high fidelity, which is helpful for tasks such as data enhancement and image style transfer. It can be used for image generation, image style transfer, data expansion, etc. For example, generate images in the style of non-existent art paintings, or generate new images for categories with less training data to expand the data set. Support Vector Machine (SVM) classifies data by mapping data into a high-dimensional space and finding the optimal hyperplane that can separate different categories of data to maximize the classification interval. Support vector machines have good performance and generalization ability in small sample and nonlinear data classification problems, and are insensitive to feature selection. They can be used for tasks such as image classification and target detection, especially when the number of samples is relatively small. For example, in some specific medical image classification tasks, they are used to distinguish normal tissue from diseased tissue.Random Forest is an ensemble learning algorithm that builds multiple decision trees and combines their prediction results (usually voting) to make the final decision. It has strong noise resistance and high accuracy, is not prone to overfitting, has strong adaptability to data, and can process high-dimensional data. It can be used for image classification, feature selection, etc. For example, when classifying images of complex scenes, using random forests can combine the judgment results of multiple decision trees to improve the accuracy of classification.

[0036] In this embodiment, optionally, using an AI algorithm to identify the test area image to obtain a probe mark detection result in the test area image includes:

[0037] Identify the test area image using an image recognition AI algorithm and / or a template matching AI algorithm;

[0038] A probe mark detection result in the test area image is determined based on the identified parameters, wherein the probe mark detection result includes a position, a shape, and a quality of the probe mark.

[0039] Among them, the test area image can be identified by an image recognition AI algorithm, a template matching AI algorithm, or a combination of the two algorithms to obtain the corresponding parameters. For example, through identification, the probability value of each pixel in the test area image being a probe mark can be obtained, and the pixel with a probability value exceeding 60% is displayed as a probe mark, and the pixel is jointly identified with the surrounding pixels to obtain the size, shape and other information of the probe marked area.

[0040] Based on the identified parameters, the probe mark detection result in the test area image can be determined, that is, whether there is a probe mark, as well as the position, shape and quality of the probe mark. The quality can be evaluated based on the size and depth of the probe mark. Since the probe mark is formed when the probe needs to be electrically connected to the pad, if the probe mark is small, shallow or does not correspond to the flat part of the pad, it can be determined that the quality is poor.

[0041] Through such a setting, the present scheme can improve the recognition accuracy of whether the image of the test area contains probe marks and the specific parameters of the probe marks using the AI ​​recognition algorithm.

[0042] In the above embodiment, optionally, the image of the test area is recognized by using an image recognition AI algorithm, including:

[0043] Using an image recognition AI algorithm to identify image features of the test area image;

[0044] The image recognition AI algorithm outputs a probability value of a suspected probe marker present in the test area image based on the image features;

[0045] or,

[0046] Using a template matching AI algorithm to identify the similarity between the suspected probe mark in the test area image and at least two pre-set standard templates;

[0047] The probability value of the suspected probe marker is determined according to the similarity.

[0048] Among them, CNN can be used for feature extraction. The preprocessed image can be convolved using the selected CNN model, and local features of the image, such as edges, textures, colors, etc., can be automatically extracted through different convolution kernels. For example, in the first convolution layer, basic edge information of the image may be extracted; as the network level deepens, more complex feature combinations are gradually extracted. Feature extraction based on other algorithms can also be performed. If SVM is used, a feature extractor needs to be manually designed, such as extracting the color histogram and shape features of the image as the input feature vector of the SVM; for GAN, the generator will also learn the feature distribution of the image in the process of generating the image.

[0049] After the features are extracted, the probability value of the suspected probe mark in the test area image can be determined based on the features, for example, a certain range is circled and the probability value that the pattern in the range is the probe mark is marked.

[0050] When using the template matching AI algorithm, it is first necessary to obtain standard templates and make at least two different standard templates based on the known probe marking features. These templates can be extracted from ideal probe marking images or generated according to design specifications. In addition, the templates are preprocessed and the standard templates are also preprocessed in the same way as the test images, such as grayscale and enhancement, to ensure matching under the same image conditions.

[0051] Template matching based on correlation coefficient calculates the correlation coefficient between the suspected probe mark in the test image and the standard template. The higher the correlation coefficient, the higher the similarity. This method is robust to illumination changes and linear transformations of the image. Template matching based on normalized cross correlation (NCC), the NCC algorithm measures the similarity by calculating the normalized cross correlation value of two image regions. Its value range is between -1 and 1, 1 indicates a perfect match, -1 indicates a complete mismatch, and 0 indicates no linear correlation.

[0052] Specifically, the sliding window method can be used to center the suspected probe mark on the test image, use a sliding window of the same size as the standard template, move the window pixel by pixel, and calculate the similarity between the image in the window and the standard template. Alternatively, for larger test images, a preliminary match can be performed at a low resolution, and after finding a roughly matching area, an accurate match can be performed at a high resolution to improve matching efficiency.

[0053] For each standard template, the similarity value between the standard template and the suspected probe mark in the test image is calculated according to the selected template matching algorithm. For example, when a template matching algorithm based on correlation coefficient is used, the correlation coefficient is calculated according to the formula as the similarity value.

[0054] The similarity values ​​of each suspected probe marker and all standard templates are recorded to form a similarity matrix.

[0055] It should be noted that it is necessary to set a similarity threshold first, and set a similarity threshold according to actual application requirements and experience. If the similarity value exceeds the threshold, it is considered that the suspected probe marker matches the corresponding standard template successfully.

[0056] According to the size of the similarity value and the threshold judgment, the standard template that the suspected probe marker in the test area image is most likely to match is determined. If multiple suspected probe markers successfully match the same standard template and the similarity is high, it can be further determined that these suspected markers may belong to the same type of probe markers.

[0057] If a suspected probe marker has a high similarity with multiple standard templates (exceeding a threshold), it is necessary to further analyze the characteristics and application scenarios of these templates, or combine other information (such as the location of the probe marker, the surrounding environment, etc.) to determine the most accurate match.

[0058] By providing two image recognition methods, this embodiment can better utilize the AI ​​algorithm to accurately identify the probe marks in the test area image, thereby providing a data basis for subsequent detection of probe abnormalities.

[0059] In the above embodiment, optionally, the test area image is identified using an image recognition AI algorithm and a template matching AI algorithm, including:

[0060] Identify image features of the test area image using an image recognition AI algorithm, and output a first probability value of a suspected probe marker present in the test area image according to the image features;

[0061] Using a template matching AI algorithm to identify the similarity between the suspected probe mark in the test area image and at least two pre-set standard templates, and determining a second probability value of the suspected probe mark according to the similarity;

[0062] The first probability value and the second probability value are added pixel by pixel to obtain a final probability value of each pixel being a probe marker.

[0063] As another solution, the above two AI algorithms can be used simultaneously, and after obtaining the first probability value and the second probability value respectively, they are added based on the weights of the two AI algorithms to obtain the final probability value of each pixel being a probe marker.

[0064] This arrangement of the present scheme can avoid the situation where the use of a single AI algorithm for identification is not accurate enough. By using multiple algorithms, the probe marker can be identified more accurately and the specific parameters of the probe marker can be identified.

[0065] Step S103, evaluating the probe mark detection result using a preset detection standard to determine whether the probe mark has a defect;

[0066] Evaluate the probe labeling test results according to pre-set criteria. Check for defects, abnormalities, or poor probe labeling and record them.

[0067] Step S104, when there is a defect in the probe mark, a processing decision is determined according to the defect type.

[0068] Among them, based on the probe marking test results and reports, corresponding decisions are made in real time and any necessary corrective measures are followed up.

[0069] This solution uses an AI algorithm real-time flaw detection system, combining AI capabilities such as deep learning and traditional image algorithms, which greatly improves the accuracy of probe performance detection. Secondly, the probe's test speed on the wafer is very fast. The AI ​​algorithm of this system is combined with a template matching algorithm to improve the detection rate, and can quickly detect problems that occur when the probe detects the wafer.

[0070] When there is a deviation in the probe performance, such as the probe mark is too close to the edge or the pad is contaminated, the AI ​​algorithm can be used to timely detect and output the results, effectively reducing probe consumption, improving detection quality, and increasing chip yield.

[0071] In the solution provided in the embodiment of the present application, a test area image is obtained, and the test area image is identified using an AI algorithm to obtain a probe mark detection result in the test area image, and the probe mark detection result is evaluated using a preset detection standard to determine whether the probe mark has a defect, so that when the probe mark has a defect, a processing decision is determined based on the defect type. By adopting this solution, it is possible to detect whether the probe has a defect based on the probe mark, so that the probe defect can be quickly and accurately detected through image recognition technology, thereby improving the detection efficiency of the probe, and it is an automated implementation process without the need to invest more human resources.

[0072] In the above embodiment, optionally, the defect type includes at least one of a probe mark being too large, a probe mark being too small, a probe mark being missing, a probe mark being offset, a probe mark being close to a pad edge, and a pad contacting the probe being contaminated.

[0073] If the probe mark is too large, the probe and the chip are easily damaged. If the probe mark is too small, the connection between the probe and the chip pad will not be tight enough, and inaccurate data may occur during the use of the probe to test the chip. The missing probe mark means that there is no electrical connection between the probe and the pad. The offset of the probe mark position may cause the pad that was originally connected to the current pad to be actually connected to other pads, resulting in errors in the process of using the probe to test the chip. If the probe mark is close to the edge of the pad, the electrical connection between the two may be unstable, which will affect the test results during the chip testing process and cause incorrect calculations of the product yield. If the pad in contact with the probe is contaminated, it may cause performance problems with the chip on the wafer, which needs to be scrapped or further processed before it can continue to be used.

[0074] By providing the above-mentioned types of anomalies, this solution can improve the accuracy of testing when using probes to test wafers, and can identify the anomalies existing in the probes in real time, thereby avoiding damage to the chip or the probe, or misidentification of chip performance.

[0075] Embodiment 2

[0076] Figure 2 It is a flow chart of the real-time flaw detection method using AI algorithm provided in the second embodiment of the present application. The specific optimization is as follows: the method further includes: determining the setting position of the microscope and the high-resolution camera, and fixing the microscope and the high-resolution camera; wherein the microscope and the high-resolution camera are set correspondingly, and at least one group is set to obtain the image of the test area from at least one angle; calibrating the microscope and the high-resolution camera.

[0077] like Figure 2 As shown, the specific steps include:

[0078] Step S201, determining the setting positions of the microscope and the high-resolution camera, and fixing the microscope and the high-resolution camera;

[0079] Wherein, the microscope and the high-resolution camera are arranged correspondingly, and at least one group is arranged to obtain an image of the test area from at least one angle;

[0080] The camera used for collecting images of the test area may be a high-resolution camera, and the microscope may be arranged in front of the high-resolution camera to collect images of the test area of ​​the chip by the probe.

[0081] Moreover, the microscope and the high-resolution camera can be set up in one group or in multiple groups, so that the test area can be photographed from different angles to obtain images of the test area at different angles.

[0082] Step S202, calibrating the microscope and the high-resolution camera;

[0083] After setting up and calibrating the probe marking detection equipment, such as the calibration of the microscope, high-resolution camera and image analysis software, the test area image can be accurately captured and identified.

[0084] Among them, microscope calibration can include calibration of the optical system, such as adjusting the light source, turning on the microscope light source, and adjusting the aperture to the maximum. By adjusting the brightness of the light source and the focus knob, the light can be evenly illuminated in the field of view. Condenser adjustment, adjust the height and aperture size of the condenser to focus the light on the sample plane. The condenser adjustment can be judged by observing the light spot on the transparent object (such as a slide) on the stage. When the light spot is uniform and fills the entire field of view, the condenser adjustment is completed. Objective lens calibration, select objective lenses of different magnifications, and after replacing the objective lens, make the image clear by fine-tuning the focus knob, and check the parfocality between the various objective lenses. If the image cannot be quickly clear after switching the objective lens, or the focus needs to be adjusted significantly, parfocality adjustment is required, which is usually performed by professional technicians. Use a standard resolution plate (such as the USAF resolution plate) placed on the stage and select a suitable objective and eyepiece combination. By observing the line pattern on the resolution plate, adjust the focus and lighting conditions of the microscope until the smallest line group can be clearly distinguished. According to the specifications of the resolution board, the resolution of the microscope at this time can be calculated to ensure that the resolution meets the design indicators of the microscope.

[0085] In addition, mechanical system calibration is also required, such as stage calibration. Use a high-precision measuring tool (such as a micrometer) to measure the difference between the actual distance the stage moves in the X and Y directions and the microscope reading. If there is a deviation, it can be calibrated by adjusting the stage's transmission mechanism or software parameters to make the actual distance the stage moves match the displayed coordinate value. Place a cross mark on the stage, rotate the stage, and observe the movement trajectory of the cross mark. If the center of rotation is not accurate, the cross mark will deviate from its original position. The center of rotation can be calibrated by adjusting the stage's rotation mechanism to ensure that the object always rotates around the same center during rotation.

[0086] For high-resolution camera calibration, internal parameter calibration can be included, such as focal length calibration. If the camera uses a fixed-focus lens, the actual focal length of the lens needs to be determined. The focal length of the lens can be calculated based on the principles of geometric optics by photographing distant objects (such as buildings or mountains) and measuring the ratio of the size of the object in the image to the actual size. For zoom lenses, focal length calibration needs to be performed at different zoom positions. Record the focal length value of the lens at each zoom position and compare it with the focal length value displayed by the camera. If there is a deviation, it can be adjusted in the camera settings. Distortion calibration, photographing a calibration plate: Use a professional chessboard or circular dot array calibration plate and place it in front of the camera to ensure that the calibration plate fills the entire field of view. Take multiple images of the calibration plate from different angles and distances. Software analysis can be used. Use camera calibration software (such as the camera calibration module in OpenCV) to import the captured calibration plate image. The software will automatically calculate the camera's distortion parameters, including radial distortion and tangential distortion, based on the feature points of the calibration plate in the image. According to the distortion parameters calculated by the software, make corrections in the camera settings, or use these parameters to dedistort the image in subsequent image processing. External parameter calibration (multi-camera system) can include relative position calibration. First, mark point setting is performed. In a multi-camera system, common mark points (such as light-emitting diodes or specific geometric marks) are set within the field of view of each camera. Coordinate measurement, by measuring the image coordinates of the mark points captured by each camera, uses the triangulation principle or other spatial coordinate calculation methods to calculate the relative position and attitude parameters between each camera. Parameter adjustment, according to the calculated relative position and attitude parameters, adjust the multi-camera system so that the spatial position relationship of each camera meets the actual needs. For example, in a stereo vision system, ensure that the optical axes of the left and right cameras are parallel and the baseline distance is accurate.

[0087] Step S203, after testing the wafer using the probe, acquiring a test area image;

[0088] Step S204, using an AI algorithm to identify the test area image to obtain a probe mark detection result in the test area image;

[0089] Step S205, evaluating the probe mark detection result using a preset detection standard to determine whether the probe mark has a defect;

[0090] Step S206, when there is a defect in the probe mark, a processing decision is determined according to the defect type.

[0091] Based on the above embodiments, this embodiment provides a solution for setting and calibrating a microscope and a high-resolution camera. Through such a setting, an image of the test area can be captured to accurately identify the probe markers, thereby improving the recognition accuracy.

[0092] In addition, some embodiments of the present application also provide an electronic device. The electronic device may be a digital computer in various forms, such as a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, etc. The electronic device may also be a mobile device in various forms, such as a personal digital assistant, a cellular phone, a smart phone, a wearable device, and other similar computing devices.

[0093] The electronic device includes: one or more processors; and a memory storing computer program instructions, wherein when the computer program instructions are executed, the processor executes the steps of the method provided in any one or more of the above embodiments. Figure 3 An exemplary structural diagram of the electronic device is disclosed. Figure 3 As shown, the electronic device includes: one or more processors 301, memory 302, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some other embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Among them, the components shown in this article, their connections and relationships, and their functions are only used as examples, and are not intended to limit the implementation of the present application described and / or required herein.

[0094] The electronic device may further include: an input device 303 and an output device 304. The processor 301, the memory 302, the input device 303 and the output device 304 may be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0095] The input device 303 can receive input digital or character information, and generate key signal input related to the user settings and function control of the electronic device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick and other input devices. The output device 304 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The display device may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display and a plasma display. In some embodiments, the display device may be a touch screen.

[0096] To provide interaction with a user, the electronic device may be a computer. The computer has: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball), through which the user can provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0097] In the embodiments of the present application, a computer program / instruction is stored on a computer-readable medium, and when the computer program / instruction is executed by a processor, the steps of the method provided by any one or more of the above embodiments are implemented. The computer-readable medium may be included in the electronic device described in the above embodiments; or it may exist independently without being assembled into the device. The above computer-readable medium carries one or more computer-readable instructions.

[0098] The memory 302 can be used as a non-transient computer-readable storage medium, which can be used to store non-transient software programs, non-transient computer executable programs and modules. The processor 301 executes various functional applications and data processing of the server by running the non-transient software programs, instructions and modules stored in the memory 302, so as to implement the program instructions / modules corresponding to the method provided by any one or more of the above embodiments in the embodiments of the present application.

[0099] The memory 302 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 302 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some embodiments, the memory 302 may optionally include a memory remotely arranged relative to the processor 301, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0100] It should be noted that the computer-readable medium described in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device.

[0101] Computer readable media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, modules of programs or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0102] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet).

[0103] In the above-described embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. For example, it can be implemented by using an application specific integrated circuit (ASIC), a general-purpose computer or any other similar hardware device. In certain embodiments, the software program of the present application can be executed by a processor to implement the above steps or functions. Similarly, the software program of the present application (including relevant data structures) can be stored in a computer-readable recording medium, for example, a RAM memory, a magnetic or optical drive or a floppy disk and similar devices. In addition, some steps or functions of the present application can be implemented by hardware, for example, as a circuit that cooperates with a processor to perform each step or function.

[0104] The computer program product provided in the embodiment of the present application includes one or more computer programs / instructions, and when the computer program / instructions are executed by the processor, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive so lid state di sk (SSD)), etc.

[0105] The flow chart or block diagram in the accompanying drawings shows the possible architecture, function and operation of the equipment, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated system for hardware that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0106] The scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and scope of the equivalent elements of the claims are included in the present application. Any figure mark in the claims should not be regarded as limiting the claims involved. In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the device claim may also be implemented by one unit or device through software or hardware. The words "first", "second", etc. are only used to distinguish the description, and do not indicate any particular order, nor can they be understood as indicating or implying relative importance.

[0107] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily mention changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims, and the above embodiments should be regarded as exemplary and non-restrictive.

Claims

1. A real-time flaw detection method using an AI algorithm, characterized in that: include: After testing the wafer using the probe, acquiring a test area image; Using an AI algorithm to identify the test area image, and obtain a probe marking detection result in the test area image; Using a preset detection standard to evaluate the probe mark detection result to determine whether the probe mark has a defect; In the event that the probe marks a defect, determining a processing decision based on the defect type; The method of using an AI algorithm to identify the test area image and obtain a probe mark detection result in the test area image includes: Identify the test area image using an image recognition AI algorithm and / or a template matching AI algorithm; The probe mark detection result in the test area image is determined based on the identified parameters, wherein the probe mark detection result includes the position, shape and quality of the probe mark, and the quality is evaluated based on the size and depth of the probe mark.

2. The method according to claim 1, characterized in that After testing the wafer using the probe and before acquiring the test area image, the method further includes: Determine the setting positions of the microscope and the high-resolution camera, and fix the microscope and the high-resolution camera; wherein the microscope and the high-resolution camera are set correspondingly, and at least one group is set to obtain an image of the test area from at least one angle; The microscope and the high-resolution camera are calibrated.

3. The method according to claim 1, characterized in that The image of the test area is identified using an image recognition AI algorithm, including: Using an image recognition AI algorithm to identify image features of the test area image; The image recognition AI algorithm outputs a probability value of a suspected probe marker present in the test area image based on the image features; or, Using a template matching AI algorithm to identify the similarity between the suspected probe mark in the test area image and at least two pre-set standard templates; The probability value of the suspected probe marker is determined according to the similarity.

4. The method according to claim 1, characterized in that: The test area image is identified using an image recognition AI algorithm and a template matching AI algorithm, including: Identify image features of the test area image using an image recognition AI algorithm, and output a first probability value of a suspected probe marker present in the test area image according to the image features; Using a template matching AI algorithm to identify the similarity between the suspected probe mark in the test area image and at least two pre-set standard templates, and determining a second probability value of the suspected probe mark according to the similarity; The first probability value and the second probability value are added pixel by pixel to obtain a final probability value of each pixel being a probe marker.

5. The method according to claim 1, characterized in that Use probes to test wafers, including: Control the probe to contact the pad surface; Controlling movement of a preset distance when the probe contacts the pad surface; Probe marking is formed by contact and movement.

6. The method according to claim 1, characterized in that The defect type includes at least one of a probe mark being too large, a probe mark being too small, a probe mark being missing, a probe mark being shifted, a probe mark being close to a pad edge, and a pad contacting the probe being contaminated.

7. An electronic device, characterized in that: The electronic device comprises: one or more processors; and A memory storing computer program instructions, which, when executed, cause the processor to perform the steps of the method as claimed in any one of claims 1 to 6.

8. A computer readable medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images