An adaptive detection method and system for a probe card aperture

Abstract

The application discloses a kind of self-adapting detection method and system of probe card aperture, it is related to aperture detection field.The method comprises: obtaining target hole image;Obtain target hole cleaning verification result;When the target hole cleaning verification result is passed, obtain real-time stage control data and real-time target positioning data;According to multistage detection feature check constraint, obtain detection feature verification result;When the detection feature verification result is qualified, obtain detection equipment control decision;Based on adaptive aperture detection algorithm, the target probe card is detected, and target aperture detection report is obtained.The method solves the technical problems that the existing probe card aperture detection method has high detection error, poor precision, and is difficult to meet the high-precision requirements of probe card aperture detection, improves the aperture detection accuracy of the probe card, and reduces the aperture detection error of the probe card.

Description

Technical Field

[0001] The present application relates to the field of aperture detection technology, and in particular to the field of aperture adaptive detection, and specifically to a method and system for adaptive detection of the aperture of a probe card. Background Art

[0002] With the rapid development of modern technology, the demand for process precision in microelectronics manufacturing is increasing. Probe cards, as a key tool in the semiconductor device manufacturing process, are widely used in wafer testing, chip packaging, and other processes. Probe card aperture inspection, a key step in ensuring stable probe card performance and improving test accuracy, has garnered widespread attention in recent years.

[0003] In summary, the present method solves the technical problems of the existing probe card aperture detection method, which has high detection error and poor precision and is difficult to meet the high-precision requirements for probe card aperture detection. Summary of the Invention

[0004] Based on this, it is necessary to provide an adaptive detection method and system for the aperture of a probe card that can improve the aperture detection accuracy of the probe card and reduce the aperture detection error of the probe card in order to address the above technical problems.

[0005] In a first aspect, a method for adaptively detecting the aperture of a probe card is provided, the method comprising: loading specification model attribute information and aperture detection accuracy parameters of a target probe card; acquiring in-aperture image data of the target probe card using an image acquisition device to obtain an in-aperture image acquisition result, and performing image enhancement on the in-aperture image acquisition result according to a Retinex algorithm to obtain an in-aperture image of a target; performing in-aperture cleaning verification on the target probe card based on the in-aperture image of the target based on a probe card in-aperture cleaning detector to obtain an in-aperture cleaning verification result of the target; and placing the target probe card on a stage of a probe card aperture detection device when the in-aperture cleaning verification result is passed. Real-time stage control data and real-time target positioning data are obtained; based on the specification model attribute information and in accordance with multi-level detection feature verification constraints, detection feature verification is performed on the real-time stage control data and the real-time target positioning data to obtain a detection feature verification result; when the detection feature verification result is qualified, the real-time stage control data and the real-time target positioning data are input into a pre-built detection equipment control decision model to obtain a detection equipment control decision; based on an adaptive aperture detection algorithm, the target probe card is detected according to the aperture detection accuracy parameters, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report.

[0006] In a second aspect, an adaptive detection system for the aperture of a probe card is provided, the system comprising: an attribute information loading module, the attribute information loading module being used to load the specification model attribute information and aperture detection accuracy parameters of the target probe card; a target in-hole image acquisition module, the target in-hole image acquisition module being used to acquire in-hole image data of the target probe card according to an image acquisition device, obtain an in-hole image acquisition result, and perform image enhancement on the in-hole image acquisition result according to a Retinex algorithm to obtain an in-hole image of the target; a target in-hole cleaning verification module, the target in-hole cleaning verification module being used to perform in-hole cleaning verification on the target probe card according to the target in-hole image based on an in-hole cleaning detector of the probe card, and obtain an in-hole cleaning verification result of the target; a control positioning data acquisition module, the control positioning data acquisition module being used to place the target probe card in the probe card aperture detection area when the in-hole cleaning verification result is passed. Real-time stage control data and real-time target positioning data are obtained on the stage of the device; a detection result acquisition module is used to obtain the detection result to be verified, and the detection feature verification of the real-time stage control data and the real-time target positioning data is performed based on the specification model attribute information and according to the multi-level detection feature verification constraint to obtain the detection feature verification result; a detection equipment control decision module is used to input the real-time stage control data and the real-time target positioning data into a pre-built detection equipment control decision model when the detection feature verification result is qualified to obtain the detection equipment control decision; a target aperture detection report acquisition module is used to detect the target probe card based on the adaptive aperture detection algorithm according to the aperture detection accuracy parameter, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report.

[0007] According to a third aspect, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps according to the first aspect when executing the computer program.

[0008] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps described in the first aspect are implemented.

[0009] The above-mentioned adaptive detection method and system for the probe card aperture adopts this method to solve the technical problems of the existing probe card aperture detection method, such as high detection error and poor accuracy, which are difficult to meet the high-precision requirements of probe card aperture detection, and achieves the technical effect of improving the aperture detection accuracy of the probe card and reducing the aperture detection error of the probe card.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 1 is a flow chart of a method for adaptively detecting the aperture of a probe card according to an embodiment;

[0012] Figure 2 A schematic diagram of a flow chart of adjusting an abnormality detection device in a method for adaptively detecting the aperture of a probe card in one embodiment;

[0013] Figure 3 is a structural block diagram of a probe card aperture adaptive detection system in one embodiment;

[0014] Figure 4 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment.

[0015] Explanation of the reference numerals: attribute information loading module 11, target hole image acquisition module 12, target hole cleaning verification module 13, control positioning data acquisition module 14, detection result acquisition module for verification 15, detection equipment control decision module 16, target aperture detection report acquisition module 17. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0017] like Figure 1 As shown, the present application provides a method for adaptively detecting the aperture of a probe card, the method comprising:

[0018] Load the target probe card's specification, model, attribute information, and aperture detection accuracy parameters;

[0019] The probe card, also known as the wafer probe card, is the key interface between the tester and the chip under test in wafer testing. It is widely used in wafer testing of technology products such as memory, logic, consumer, drive, and communication ICs. It is mainly composed of PCB, probes, and functional components. Depending on different needs, it also includes electronic components, reinforcement plates, and other components. During wafer testing, the tester contacts the wafer through the probe card and transmits test signals to obtain the electrical signals of the wafer. The principle of using the probe card is to directly contact the probes on it with the pads or bumps on the chip, derive the chip signals, and then cooperate with peripheral test instruments and software control to achieve automated measurement of the wafer. The present application provides an adaptive detection method for the aperture of a probe card, which improves the aperture detection accuracy of the probe card and reduces the aperture detection error of the probe card.

[0020] The target probe card refers to the probe card arbitrarily selected by the staff for research in this application, recorded as the target probe card, and the specification and model attribute information and aperture detection accuracy parameters of the target probe card are obtained, wherein the specification and model attribute information includes the size, material, number of probes and probe layout of the target probe card, and the physical properties of the target probe card are obtained. The aperture detection accuracy parameter refers to the minimum resolution or maximum error that the probe card can achieve when performing aperture detection, which is used to ensure the accuracy and reliability of the test; the specification and model attribute information and aperture detection accuracy parameters of the target probe card are obtained by referring to relevant product manuals, technical specifications, etc.; data support is provided for the subsequent improvement of the aperture detection accuracy of the target probe card.

[0021] Acquire the in-hole image data of the target probe card using an image acquisition device to obtain an in-hole image acquisition result, and perform image enhancement on the in-hole image acquisition result according to a Retinex algorithm to obtain an in-hole image of the target;

[0022] An image acquisition device refers to an instrument for acquiring images, such as a microscope camera, etc. The image acquisition device is used to photograph the in-hole area of ​​the target probe card to obtain in-hole image data, i.e., the in-hole image acquisition result. The Retinex algorithm is an image enhancement algorithm based on color constancy, which is used to improve the contrast and brightness distribution of the image. The in-hole image acquisition result is processed by the Retinex algorithm, and the in-hole image acquisition result is enhanced by estimating and removing the illumination component in the image. The enhanced in-hole image acquisition result is obtained as the target in-hole image. The image acquisition device and the Retinex algorithm are used to effectively acquire and enhance the in-hole image of the target probe card, so as to more clearly observe and analyze the structure and details in the hole.

[0023] Based on the probe card in-hole cleaning detector, performing in-hole cleaning verification on the target probe card according to the target in-hole image to obtain a target in-hole cleaning verification result;

[0024] The probe card hole cleaning detector includes a probe card hole cleaning evaluation unit and a probe card hole cleaning verification unit;

[0025] Inputting the target hole interior image into the probe card hole interior cleaning evaluation unit to obtain the target hole interior cleaning evaluation coefficient;

[0026] Inputting the target hole cleaning evaluation coefficient into the probe card hole cleaning verification unit to generate the target hole cleaning verification result;

[0027] The probe card hole cleaning verification unit includes a probe card hole cleaning verification constraint rule, wherein the probe card hole cleaning verification constraint rule includes that when the target hole cleaning evaluation coefficient satisfies the probe card hole cleaning evaluation constraint, the target hole cleaning verification result obtained is passed;

[0028] The probe card hole cleaning verification constraint rule further includes obtaining a target hole cleaning verification result as a failure when the target hole cleaning evaluation coefficient does not meet the probe card hole cleaning evaluation constraint, and generating a probe card hole cleaning optimization instruction.

[0029] The probe card in-hole cleaning detector is a module for detecting the cleanliness of the in-hole area of ​​the target probe card and evaluating the cleanliness of the in-hole, including a probe card in-hole cleaning evaluation unit and a probe card in-hole cleaning verification unit. The probe card in-hole cleaning evaluation unit is used to analyze the input target in-hole image and extract features related to the cleanliness of the in-hole, such as the number, size, distribution of particulate matter in the hole, etc. According to the embedded calculation method of the probe card in-hole cleaning evaluation unit, the cleanliness of the in-hole of the target probe card, that is, the cleanliness evaluation coefficient, is obtained. The significance of the cleaning verification is that the cleanliness of the sample directly affects the accuracy of the measurement result. If there are impurities or residues in the aperture, the measurement value may be too large or too small; the probe card in-hole cleaning verification unit has a built-in probe card in-hole cleaning verification constraint rule, and the probe card in-hole cleaning verification constraint rule defines the qualified standard for the cleanliness of the in-hole. If the cleaning evaluation coefficient meets the cleaning evaluation constraint, that is, the cleanliness of the hole reaches the qualified standard, the verification unit will generate a target in-hole cleaning verification result of "pass". If the cleaning evaluation coefficient does not meet the cleaning evaluation constraint, that is, there are non-standard contaminants in the hole, the verification unit will generate a target hole cleaning verification result of "fail", that is, the probe card hole cleaning evaluation constraint. In the probe card hole cleaning verification unit, the probe card hole cleaning evaluation constraint is used to compare with the cleaning evaluation coefficient. When the target hole cleaning verification result is "fail", a probe card hole cleaning optimization instruction is generated at the same time. The probe card hole cleaning optimization instruction includes cleaning methods, etc. The comparison results include "pass" and "fail", both of which are called the target hole cleaning verification result. The target hole image is input into the probe card hole cleaning evaluation unit to obtain the target hole cleaning evaluation coefficient. The target hole cleaning evaluation coefficient is input into the probe card hole cleaning verification unit. Verification is performed according to the built-in cleaning verification constraint rules to obtain a verification result. Based on the verification result, the target hole cleaning verification result is obtained, and a cleaning optimization instruction is generated as needed. This can efficiently and accurately verify the cleanliness of the probe card hole, ensuring that the probe card meets the required cleaning standards before use.

[0030] Based on big data, load multiple probe card hole cleaning evaluation record data sets;

[0031] Based on multiple image feature learning operators, the multiple probe card in-hole cleaning evaluation record data sets are trained and tested respectively to obtain multiple in-hole cleaning evaluation models that meet preset in-hole cleaning evaluation accuracy constraints;

[0032] Obtaining a plurality of in-hole cleaning evaluation accuracy parameters corresponding to the plurality of in-hole cleaning evaluation models, and performing a proportion calculation based on the plurality of in-hole cleaning evaluation accuracy parameters to construct an in-hole cleaning evaluation output incentive model, wherein the in-hole cleaning evaluation output incentive model includes a plurality of in-hole cleaning evaluation output incentive coefficients;

[0033] embedding the plurality of in-hole cleaning evaluation models and the in-hole cleaning evaluation output excitation model into the in-hole cleaning evaluation unit of the probe card;

[0034] Inputting the target hole image into the multiple hole cleaning evaluation models to obtain multiple hole cleaning evaluation coefficients;

[0035] The multiple in-hole cleaning evaluation coefficients are input into the in-hole cleaning evaluation output excitation model, and the multiple in-hole cleaning evaluation coefficients are weighted calculated according to the multiple in-hole cleaning evaluation output excitation coefficients to generate the target in-hole cleaning evaluation coefficient.

[0036] Based on big data, a plurality of probe card in-hole cleaning evaluation record data sets are obtained. The probe card in-hole cleaning evaluation record data sets refer to a data set formed by integrating the record data of the in-hole cleaning evaluation of a plurality of probe cards in the past time period, and are used to train a plurality of subsequent in-hole cleaning evaluation models. According to the characteristics of the probe card in-hole images, a plurality of operators suitable for image feature learning are selected, such as convolutional neural networks, autoencoders, and other machine learning algorithms suitable for image recognition. The plurality of probe card in-hole cleaning evaluation record data sets are trained according to a plurality of different machine learning algorithms. By adjusting model parameters and optimizing algorithms, the plurality of in-hole cleaning evaluation models are enabled to accurately identify the in-hole cleanliness and output a plurality of in-hole cleaning evaluation accuracy parameters. For example, the target in-hole images are input into the plurality of in-hole cleaning evaluation models, and a plurality of different in-hole cleaning evaluation coefficients obtained by the models trained with different algorithms for the target in-hole images are obtained; a plurality of in-hole cleaning evaluation accuracy parameters corresponding to the plurality of in-hole cleaning evaluation models are obtained, i.e., the model accuracy of the plurality of in-hole cleaning evaluation models is obtained, wherein the model accuracy refers to the model accuracy rate, which is an indicator for measuring the ability of the machine learning model to make predictions or decisions based on data. For example, using the plurality of probe card in-hole cleaning evaluation record data sets The in-hole cleaning evaluation model is trained, and the output result of the in-hole cleaning evaluation model conforms to the ratio of the number of the plurality of probe card in-hole cleaning evaluation record data sets to the number of the plurality of probe card in-hole cleaning evaluation record data sets, which is the in-hole cleaning evaluation accuracy parameter; and a proportion calculation is performed based on the plurality of in-hole cleaning evaluation accuracy parameters to construct an in-hole cleaning evaluation output incentive model, wherein the in-hole cleaning evaluation output incentive model includes a plurality of in-hole cleaning evaluation output incentive coefficients, and the plurality of in-hole cleaning evaluation output incentive coefficients refer to the ratio of the in-hole cleaning evaluation accuracy parameter to the sum of the plurality of in-hole cleaning evaluation accuracy parameters. value, embed the multiple in-hole cleaning evaluation models and the in-hole cleaning evaluation output excitation model into the probe card in-hole cleaning evaluation unit; input the multiple in-hole cleaning evaluation coefficients into the in-hole cleaning evaluation output excitation model, and perform weighted calculation on the multiple in-hole cleaning evaluation coefficients according to the multiple in-hole cleaning evaluation output excitation coefficients to generate the target in-hole cleaning evaluation coefficient. By using big data and machine learning technology, a more accurate and reliable probe card in-hole cleaning evaluation system can be constructed. The system can comprehensively consider the outputs of multiple models and perform weighted calculation according to the performance of each model, thereby improving the accuracy and stability of the evaluation.

[0037] When the cleaning verification result in the target hole is passed, the target probe card is placed on a stage of a probe card aperture detection device to obtain real-time stage control data and real-time target positioning data;

[0038] The stage is a component specifically designed to position and support the probe card under test. The probe card aperture detection device is a device used to perform subsequent aperture measurements on the target probe card. When the target hole cleaning verification result passes, the target probe card's hole cleaning is certified, and the target probe card can proceed to the next test item. While the stage is operating, it generates real-time stage control data, recording parameters such as the stage's motion trajectory, height, and inclination, reflecting the stage's real-time status as it adjusts the probe card's position. Real-time target positioning data is acquired via high-precision sensors, capturing the probe card's image or signal in real time. This data is then processed through algorithms to determine the specific position of each hole within the synchronization timeframe. This real-time target positioning data provides critical coordinate information for subsequent aperture measurements, ensuring measurement accuracy. When the target hole cleaning verification result passes, the target probe card is placed on the stage of the probe card aperture detection device, and real-time stage control data and target positioning data are acquired, ensuring the accuracy and reliability of the aperture measurement.

[0039] Based on the specification and model attribute information, and in accordance with multi-level detection feature verification constraints, performing detection feature verification on the real-time stage control data and the real-time target positioning data to obtain a detection feature verification result;

[0040] Multi-level detection feature verification constraints are detection rules and conditions established based on specification and model attribute information. They are used to verify the real-time stage control data and real-time target positioning data. These constraints include the accuracy requirements for the stage's movement trajectory, the error range of the positioning data, and the recognition accuracy of specific hole positions. During the detection feature verification process, key features of the real-time stage control data and real-time target positioning data are extracted, such as the coordinate points of the movement trajectory and the coordinate positions of the positioning data. These key features are then compared and verified against the multi-level detection feature verification constraints. If the data features meet the verification constraint requirements, the verification is considered passed; if the data features do not meet the requirements, a corresponding error message is generated to facilitate further adjustment and optimization.

[0041] Based on the specification and model attribute information, stage control anomaly evaluation records of similar probe cards are collected to obtain a similar stage control anomaly evaluation record set, and a stage control anomaly evaluation network that satisfies a first convergence constraint is trained based on the similar stage control anomaly record set;

[0042] Collecting positioning anomaly evaluation records of the same type of probe cards based on the specification and model attribute information to obtain a similar positioning anomaly evaluation record set, and training a positioning anomaly evaluation network that satisfies a second convergence constraint based on the similar positioning anomaly evaluation record set;

[0043] Inputting the real-time stage control data into the stage control abnormality evaluation network to obtain a real-time stage control abnormality evaluation coefficient;

[0044] Inputting the real-time target positioning data into the positioning anomaly evaluation network to obtain a real-time target positioning anomaly evaluation coefficient;

[0045] Comparing the multi-level detection feature verification constraint with the real-time stage control anomaly evaluation coefficient and the real-time target positioning anomaly evaluation coefficient to generate the detection feature verification result;

[0046] Wherein, the multi-level detection feature verification constraints include a stage control abnormality evaluation threshold and a target positioning abnormality evaluation threshold;

[0047] When the real-time stage control abnormality evaluation coefficient is less than the stage control abnormality evaluation threshold, and the real-time target positioning abnormality evaluation coefficient is less than the target positioning abnormality evaluation threshold, the generated detection feature verification result is qualified;

[0048] When the real-time stage control abnormality evaluation coefficient is greater than / equal to the stage control abnormality evaluation threshold, and / or the real-time target positioning abnormality evaluation coefficient is greater than / equal to the target positioning abnormality evaluation threshold, the generated detection feature verification result is unqualified.

[0049] According to the specification and model attribute information of the target probe card, the stage control anomaly evaluation records of the same probe card are collected. The stage control anomaly evaluation record set includes various abnormal situations that occurred in the stage control of the same probe card during the detection process in the past time period, which provides a basis for the training of the stage control anomaly evaluation network. The stage control anomaly evaluation network that meets the first convergence constraint is trained through the same stage control anomaly evaluation record set. The first convergence constraint refers to the model accuracy of the stage control anomaly evaluation network. When the accuracy of the stage control anomaly evaluation network meets the first convergence constraint, the stage control anomaly evaluation network is obtained; the same positioning anomaly evaluation record set refers to various abnormal situations that occurred in the positioning data of the same probe card during the detection process. According to the above technology, based on the same positioning anomaly evaluation record set, a positioning anomaly evaluation network that meets the second convergence constraint is trained. The positioning anomaly evaluation network can learn and identify abnormal patterns of positioning data. The second convergence constraint refers to the model accuracy of the positioning anomaly evaluation network, which provides support for anomaly evaluation of real-time data. The real-time stage control data is input into the stage control anomaly evaluation network to obtain a real-time stage control anomaly evaluation coefficient; the real-time target positioning data is input into the positioning anomaly evaluation network to obtain a real-time target positioning anomaly evaluation coefficient; wherein the real-time stage control anomaly evaluation coefficient includes the real-time stage data anomaly, and similarly, the target positioning anomaly evaluation coefficient includes the position anomaly of the real-time probe card; the multi-level detection feature verification constraint is compared with the real-time stage control anomaly evaluation coefficient and the real-time target positioning anomaly evaluation coefficient to generate the detection feature verification result, wherein the multi-level detection feature verification constraint includes a stage control anomaly evaluation threshold and a target positioning anomaly evaluation threshold. If the real-time stage control anomaly evaluation coefficient is less than the stage control anomaly evaluation threshold, and the real-time target positioning anomaly evaluation coefficient is less than the target positioning anomaly evaluation threshold, then the generated detection feature verification result is qualified; otherwise, the generated detection feature verification result is unqualified. Through the above method, it can be ensured that the stage control and target positioning data meet the requirements before aperture detection, thereby improving the accuracy and reliability of detection.

[0050] When the detection feature verification result is qualified, the real-time stage control data and the real-time target positioning data are input into a pre-built detection equipment control decision model to obtain a detection equipment control decision;

[0051] The detection equipment control decision model is used to perform aperture detection on a target probe card after both the hole cleanliness and the stage have been inspected and no abnormalities have been found. The detection equipment control decision model is constructed based on a neural network model, and historical stage control data, historical target positioning data, and historical detection equipment control decisions are obtained. The historical detection equipment control decisions refer to historical parameters of the stage during past testing of the target probe card under conditions corresponding to the historical stage control data and historical target positioning data. The historical stage control data and historical target positioning data are input into the detection equipment control decision model, and the detection equipment control decision model is optimized based on the corresponding historical detection equipment control decisions. When the model accuracy of the detection equipment control decision model meets preset requirements, the detection equipment control decision model is obtained. When the detection feature verification result is qualified, the real-time stage control data and real-time target positioning data are input into the pre-constructed detection equipment control decision model, and the detection equipment control decision is output. The detection equipment control decision is the adjustment parameter of the stage, such as the inclination angle and height of the stage. The above method helps to achieve more accurate and efficient aperture detection.

[0052] like Figure 2 As shown, when the detection feature verification result is unqualified, abnormal device detection and device abnormal feature data are generated;

[0053] Performing equipment abnormality compensation analysis based on the equipment abnormality feature data to determine equipment abnormality compensation adjustment data;

[0054] The abnormality detection device is adjusted based on the device abnormality compensation adjustment data.

[0055] When the detection feature verification result is unqualified, the detection abnormality device and the device abnormality feature data are generated, the detection abnormality device includes an abnormal stage and an abnormal target probe card placed on the stage, and the device abnormality feature data includes the abnormal conditions of the stage and the abnormal conditions of the target probe card, such as abnormal fluctuations in the detection data, values ​​exceeding a preset range, specific patterns that do not conform to normal patterns, etc. It should be clear that when any of the stage and the target probe card is abnormal, the detection abnormality device and the device abnormality feature data will be generated; based on the device abnormality feature data, device abnormality compensation analysis is performed, including Statistical analysis of abnormal data, pattern recognition, and expert system-based abnormality diagnosis, through equipment abnormality compensation analysis, can determine the specific cause of the equipment abnormality and calculate the corresponding equipment abnormality compensation adjustment data. The equipment abnormality compensation adjustment data is a set of parameters used to adjust and repair the equipment status, including adjusting the sensitivity of the detection equipment and the positioning accuracy of the calibration equipment. Based on the equipment abnormality compensation adjustment data, the system adjusts the abnormality detection equipment. The adjustment process includes automatically adjusting the parameter settings of the equipment and triggering the self-calibration program of the equipment. By adjusting the abnormality detection equipment, the equipment abnormality is eliminated and restored to normal working state. When the detection feature verification result is unqualified, the system effectively responds to the equipment abnormality by generating equipment abnormality feature data, performing equipment abnormality compensation analysis, and adjusting the abnormality detection equipment based on the analysis results, ensuring the accuracy of the detection process and the stable operation of the equipment.

[0056] Based on the adaptive aperture detection algorithm, the target probe card is detected according to the aperture detection accuracy parameter, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report.

[0057] The adaptive aperture detection algorithm can set appropriate detection thresholds and accuracy requirements based on the aperture detection accuracy parameters. Combined with the detection equipment control decision, it can guide the detection equipment to perform detection according to the optimal path and strategy, including adjusting the movement trajectory of the stage, optimizing positioning accuracy, controlling the movement speed of the probe card, etc. When the probe card aperture detection device is used to detect the target probe card, the adaptive aperture detection algorithm will generate a detailed target aperture detection report, which includes the aperture measurement results and provides detailed information about the detection equipment performance, detection accuracy and abnormal conditions, which helps operators to have an in-depth understanding of the detection results and provides strong support for subsequent quality control and process improvements.

[0058] Inputting the aperture detection accuracy parameter into a pre-constructed aperture detection frequency measurement table to obtain a matching aperture detection frequency;

[0059] Taking the matching aperture detection frequency as a detection frequency constraint, controlling the probe card aperture detection device to perform multiple detections on the target probe card according to the detection equipment control decision, to obtain multiple sets of aperture detection data that meet the detection frequency constraint;

[0060] Collecting real-time environmental information and real-time equipment monitoring information corresponding to the multiple sets of aperture detection data to obtain multiple sets of aperture detection scene data;

[0061] performing adaptive detection scene anomaly compensation on the multiple sets of aperture detection data based on the multiple sets of aperture detection scene data to obtain multiple aperture detection results;

[0062] Data fusion is performed based on the multiple aperture detection results to generate the target aperture detection report.

[0063] The aperture detection accuracy parameters are input into a pre-constructed aperture detection frequency measurement table to determine the appropriate detection frequency of the target probe card for specific accuracy requirements. The aperture detection frequency measurement table is constructed based on historical data, and the historical data includes the aperture detection accuracy parameters and the corresponding detection frequencies. It can automatically match the corresponding aperture detection times according to different accuracy parameters, that is, match the aperture detection frequency, to ensure that the detection process will not miss important information due to too low a frequency, nor waste resources due to too high a frequency. Taking the matching aperture detection frequency as the detection frequency constraint, the probe card aperture detection device is controlled according to the detection equipment control decision to perform multiple detections on the target probe card. In the multiple detections, the detection equipment will operate according to the parameters and strategies specified in the detection equipment control decision to ensure that each detection can accurately reflect the aperture condition of the target probe card. Through multiple detections, multiple sets of aperture detection data that meet the detection frequency constraint can be obtained; real-time environmental information and real-time equipment monitoring information corresponding to the multiple sets of aperture detection data are collected, that is, environmental information and equipment monitoring information when the multiple sets of aperture detection data are collected, including environmental factors such as temperature, humidity, and air pressure during detection, as well as monitoring data such as the operating status, temperature, and vibration of the equipment, to constitute multiple sets of aperture detection scene data; adaptive detection scene anomaly compensation is performed on the multiple sets of aperture detection data based on the multiple sets of aperture detection scene data in order to eliminate or reduce the influence of environmental factors and equipment status on the detection results, and to obtain more accurate multiple aperture detection results by identifying and compensating for abnormal deviations caused by scene factors. Data fusion is performed based on multiple aperture detection results. Data fusion combines multiple detection results to form a more comprehensive and accurate conclusion. By considering the data differences and correlations under different detection conditions, more robust and reliable aperture detection results can be obtained, and a target aperture detection report is generated. The target aperture detection report includes detailed detection data, analysis results and conclusions. By inputting aperture detection accuracy parameters to obtain matching detection frequency, the probe card aperture detection device is controlled to perform multiple detections, real-time environmental information and equipment monitoring information are collected, adaptive detection scene anomaly compensation is performed, and finally data fusion is performed to generate a report, thereby improving the quality and efficiency of the detection process.

[0064] Traversing the multiple sets of aperture detection data and the multiple sets of aperture detection scene data, extracting a first set of aperture detection data and a first set of aperture detection scene data corresponding to the first set of aperture detection data, wherein the first set of aperture detection scene data includes first aperture detection environment information and first aperture detection device monitoring information;

[0065] Performing anomaly recognition and evaluation on the first aperture detection environment information to determine a first environment anomaly recognition result and a first environment anomaly evaluation coefficient;

[0066] Performing abnormality identification and evaluation on the monitoring information of the first aperture detection device to obtain a first device abnormality identification result and a first device abnormality evaluation coefficient;

[0067] If the first environmental anomaly evaluation coefficient is greater than / equal to the environmental anomaly evaluation threshold, and the first device anomaly evaluation coefficient is less than the device anomaly evaluation threshold, inputting the first environmental anomaly recognition result into an environmental anomaly compensation analysis model to obtain first environmental anomaly compensation data, and adaptively compensating the first set of aperture detection scene data based on the first environmental anomaly compensation data to obtain a first aperture detection result, and adding the first aperture detection result to the multiple aperture detection results;

[0068] If the first environment abnormality evaluation coefficient is less than the environment abnormality evaluation threshold, and the first device abnormality evaluation coefficient is greater than / equal to the device abnormality evaluation threshold, inputting the first device abnormality identification result into the device abnormality compensation analysis model to obtain first device abnormality compensation data, and adaptively compensating the first set of aperture detection scene data based on the first device abnormality compensation data to generate the first aperture detection result;

[0069] If the first environment abnormality evaluation coefficient is greater than / equal to the environment abnormality evaluation threshold, and the first device abnormality evaluation coefficient is greater than / equal to the device abnormality evaluation threshold, based on the first environment abnormality identification result and the first device abnormality identification result, abnormality compensation fusion is performed according to the environment abnormality compensation analysis model and the device abnormality compensation analysis model to obtain first fused abnormality compensation data, and adaptive compensation is performed on the first set of aperture detection scene data based on the first fused abnormality compensation data to obtain the first aperture detection result;

[0070] If the first environmental abnormality evaluation coefficient is less than the environmental abnormality evaluation threshold, and the first equipment abnormality evaluation coefficient is less than the equipment abnormality evaluation threshold, a first detection data mapping instruction is obtained, and the first set of aperture detection data is output as the first aperture detection result according to the first detection data mapping instruction.

[0071] Traversing the multiple groups of aperture detection data and the multiple groups of aperture detection scene data, extracting a first group of aperture detection data and a first group of aperture detection scene data corresponding to the first group of aperture detection data, wherein the first group of aperture detection scene data includes first aperture detection environment information and first aperture detection equipment monitoring information, the first group of aperture detection data refers to any group of data extracted from the multiple groups of aperture detection data for research, the first aperture detection environment information and the first aperture detection equipment monitoring information are the corresponding data of the first group of aperture detection data, performing abnormality identification and evaluation on the first aperture detection environment information refers to checking whether environmental parameters such as temperature, humidity, air pressure, etc. are within a normal range, and whether there are sudden changes or abnormal fluctuations; determining a first environmental anomaly identification result and a first environmental anomaly evaluation coefficient, the first environmental anomaly identification result refers to an identified abnormal environmental condition, the first environmental anomaly evaluation coefficient refers to an evaluation parameter for the abnormal environmental condition, constructing an environmental anomaly compensation analysis model, the environmental anomaly compensation analysis model is constructed based on a neural network model, and the input data is the first environmental anomaly identification result, The output data is the first environmental anomaly compensation data. The historical environmental anomaly recognition result data set and the historical environmental anomaly compensation data are obtained to train the environmental anomaly compensation analysis model. When the model accuracy of the environmental anomaly compensation analysis model reaches the requirement, the environmental anomaly compensation analysis model is obtained. Similarly, the equipment anomaly compensation analysis model can be constructed. If the environmental anomaly evaluation coefficient is greater than or equal to the environmental anomaly evaluation threshold, and the equipment anomaly evaluation coefficient is less than the equipment anomaly evaluation threshold, it means that the environment has a greater impact on the detection result and the equipment status is normal. At this time, the environmental anomaly recognition result is input into the environmental anomaly compensation analysis model to obtain the first environmental anomaly compensation data, and the first set of aperture detection scene data is adaptively compensated to obtain the first aperture detection result; if the environmental anomaly evaluation coefficient is less than the environmental anomaly evaluation threshold, and the equipment anomaly evaluation coefficient is greater than or equal to the equipment anomaly evaluation threshold, it means that the abnormal equipment status has a greater impact on the detection result. At this time, the equipment anomaly recognition result is input into the equipment anomaly compensation analysis model to obtain the first equipment anomaly compensation data, and corresponding adaptive compensation is performed; if both exceed their respective thresholds, it means that both the environment and the equipment have abnormal influences. At this point, based on the environmental anomaly identification results and the equipment anomaly identification results, two compensation analysis models are used to perform anomaly compensation fusion, obtaining the first fused anomaly compensation data, and then performing adaptive compensation. Anomaly compensation fusion refers to adjusting both anomalies and using the acquired parameters as the first fused anomaly compensation data. If neither exceeds their respective thresholds, indicating that the environment and equipment status are normal, no compensation is required, and the first set of aperture detection data is directly used as the first aperture detection result. The obtained aperture detection result, namely the first aperture detection result, is added to the multiple aperture detection results to facilitate subsequent data fusion and report generation.Each set of aperture detection data and the corresponding scene data are processed identically until all data has been processed, accurately reflecting the actual aperture detection situation, reducing the impact of environmental and equipment anomalies on the test results, and improving the accuracy and reliability of the report. This method solves the technical problems of existing probe card aperture detection methods, such as high detection error and poor accuracy, which make it difficult to meet the high-precision requirements of probe card aperture detection. It achieves the technical effect of improving the accuracy of probe card aperture detection and reducing the error of probe card aperture detection.

[0072] like Figure 3 As shown, an embodiment of the present application includes an adaptive detection system for a probe card aperture, the system comprising:

[0073] An attribute information loading module 11 is used to load the specification and model attribute information of the target probe card and the aperture detection accuracy parameters;

[0074] a target in-hole image acquisition module 12, configured to acquire in-hole image data of the target probe card using an image acquisition device to obtain an in-hole image acquisition result, and perform image enhancement on the in-hole image acquisition result using a Retinex algorithm to obtain a target in-hole image;

[0075] A target hole cleaning verification module 13 is configured to perform a target hole cleaning verification on the target probe card according to the target hole image based on a probe card hole cleaning detector to obtain a target hole cleaning verification result;

[0076] a control positioning data acquisition module 14 for placing the target probe card on a stage of a probe card aperture detection device to obtain real-time stage control data and real-time target positioning data when the cleaning verification result in the target hole is passed;

[0077] A detection result obtaining module 15 is configured to perform detection feature verification on the real-time stage control data and the real-time target positioning data based on the specification and model attribute information and in accordance with multi-level detection feature verification constraints to obtain a detection feature verification result;

[0078] a detection device control decision module 16, wherein the detection device control decision module 16 is configured to input the real-time stage control data and the real-time target positioning data into a pre-built detection device control decision model to obtain a detection device control decision when the detection feature verification result is qualified;

[0079] The target aperture detection report acquisition module 17 is used to detect the target probe card based on the adaptive aperture detection algorithm, the aperture detection accuracy parameters, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report.

[0080] Furthermore, the embodiment of the present application also includes:

[0081] The probe card hole cleaning detector includes a module, wherein the probe card hole cleaning detector includes a module for the probe card hole cleaning detector including a probe card hole cleaning evaluation unit and a probe card hole cleaning verification unit;

[0082] a target hole cleaning evaluation coefficient obtaining module, the target hole cleaning evaluation coefficient obtaining module being used to input the target hole interior image into the probe card hole cleaning evaluation unit to obtain the target hole cleaning evaluation coefficient;

[0083] a cleaning validation result generating module, the cleaning validation result generating module being configured to input the cleaning evaluation coefficient in the target well into the probe card well cleaning verification unit to generate the cleaning validation result in the target well;

[0084] a cleaning evaluation coefficient judgment module, wherein the cleaning evaluation coefficient judgment module is used in the probe card hole cleaning verification unit including a probe card hole cleaning verification constraint rule, the probe card hole cleaning verification constraint rule including that when the target hole cleaning evaluation coefficient satisfies the probe card hole cleaning evaluation constraint, the target hole cleaning verification result obtained is passed;

[0085] A cleaning optimization instruction generation module, which is used for the cleaning verification constraint rules in the probe card hole, also includes when the cleaning evaluation coefficient in the target hole does not meet the cleaning evaluation constraint in the probe card hole, the cleaning verification result in the target hole is obtained as failure, and a cleaning optimization instruction in the probe card hole is generated.

[0086] Furthermore, the embodiment of the present application also includes:

[0087] A cleaning evaluation record data set loading module, wherein the cleaning evaluation record data set loading module is used to load a plurality of probe card hole cleaning evaluation record data sets based on big data;

[0088] an in-hole cleaning evaluation model constraint module, the in-hole cleaning evaluation model constraint module being used to train and test the plurality of probe card in-hole cleaning evaluation record data sets based on a plurality of image feature learning operators, to obtain a plurality of in-hole cleaning evaluation models that meet preset in-hole cleaning evaluation accuracy constraints;

[0089] A cleaning evaluation output incentive module, the cleaning evaluation output incentive module is used to obtain multiple in-hole cleaning evaluation accuracy parameters corresponding to the multiple in-hole cleaning evaluation models, and perform a proportion calculation based on the multiple in-hole cleaning evaluation accuracy parameters to construct an in-hole cleaning evaluation output incentive model, wherein the in-hole cleaning evaluation output incentive model includes multiple in-hole cleaning evaluation output incentive coefficients;

[0090] a probe card hole cleaning evaluation unit module, the probe card hole cleaning evaluation unit module being used to embed the plurality of hole cleaning evaluation models and the hole cleaning evaluation output excitation model into the probe card hole cleaning evaluation unit;

[0091] an in-hole cleaning evaluation coefficient obtaining module, configured to input the target in-hole image into the plurality of in-hole cleaning evaluation models to obtain a plurality of in-hole cleaning evaluation coefficients;

[0092] The in-hole cleaning evaluation coefficient weighted calculation module is used to input the multiple in-hole cleaning evaluation coefficients into the in-hole cleaning evaluation output excitation model, perform weighted calculation on the multiple in-hole cleaning evaluation coefficients according to the multiple in-hole cleaning evaluation output excitation coefficients, and generate the target in-hole cleaning evaluation coefficient.

[0093] Furthermore, the embodiment of the present application also includes:

[0094] a stage control anomaly evaluation network module, the stage control anomaly evaluation network module being configured to collect stage control anomaly evaluation records of similar probe cards based on the specification and model attribute information, obtain a similar stage control anomaly evaluation record set, and train a stage control anomaly evaluation network that satisfies a first convergence constraint based on the similar stage control anomaly record set;

[0095] A positioning anomaly evaluation network module, the positioning anomaly evaluation network module being used to collect positioning anomaly evaluation records of the same type of probe cards based on the specification and model attribute information, obtain a similar positioning anomaly evaluation record set, and train a positioning anomaly evaluation network that satisfies a second convergence constraint based on the similar positioning anomaly evaluation record set;

[0096] a control abnormality evaluation coefficient obtaining module, the control abnormality evaluation coefficient obtaining module being used to input the real-time stage control data into the stage control abnormality evaluation network to obtain a real-time stage control abnormality evaluation coefficient;

[0097] A positioning anomaly evaluation coefficient module, wherein the positioning anomaly evaluation coefficient module is used to input the real-time target positioning data into the positioning anomaly evaluation network to obtain a real-time target positioning anomaly evaluation coefficient;

[0098] a detection feature verification result generation module, the detection feature verification result generation module being used to compare the multi-level detection feature verification constraint with the real-time stage control anomaly evaluation coefficient and the real-time target positioning anomaly evaluation coefficient to generate the detection feature verification result;

[0099] A positioning anomaly evaluation threshold acquisition module, wherein the positioning anomaly evaluation threshold acquisition module is used, wherein the multi-level detection feature verification constraint includes a stage control anomaly evaluation threshold and a target positioning anomaly evaluation threshold;

[0100] A qualified result generation module, wherein the qualified result generation module is configured to generate the detection feature verification result as qualified when the real-time stage control abnormality evaluation coefficient is less than the stage control abnormality evaluation threshold, and the real-time target positioning abnormality evaluation coefficient is less than the target positioning abnormality evaluation threshold;

[0101] An unqualified result generation module is used to generate an unqualified detection feature verification result when the real-time stage control abnormality evaluation coefficient is greater than / equal to the stage control abnormality evaluation threshold, and / or the real-time target positioning abnormality evaluation coefficient is greater than / equal to the target positioning abnormality evaluation threshold.

[0102] Furthermore, the embodiment of the present application also includes:

[0103] An abnormal data generation module, the abnormal data generation module is used to generate abnormal device detection and device abnormality feature data when the detection feature verification result is unqualified;

[0104] an abnormality compensation adjustment data module, the abnormality compensation adjustment data module being used to perform equipment abnormality compensation analysis based on the equipment abnormality feature data and determine equipment abnormality compensation adjustment data;

[0105] An abnormal device adjustment module is used to adjust the abnormal device based on the device abnormality compensation adjustment data.

[0106] Furthermore, the embodiment of the present application also includes:

[0107] A matching aperture detection frequency acquisition module, wherein the matching aperture detection frequency acquisition module is used to input the aperture detection accuracy parameter into a pre-built aperture detection frequency measurement table to obtain a matching aperture detection frequency;

[0108] a multiple detection module, the multiple detection module being configured to control the probe card aperture detection device to perform multiple detections on the target probe card based on the detection device control decision, taking the matching aperture detection frequency as a detection frequency constraint, and obtaining multiple sets of aperture detection data that satisfy the detection frequency constraint;

[0109] An aperture detection scene data acquisition module, the aperture detection scene data acquisition module is used to collect real-time environmental information and real-time equipment monitoring information corresponding to the multiple sets of aperture detection data to obtain multiple sets of aperture detection scene data;

[0110] an aperture detection result obtaining module, configured to perform adaptive detection scene anomaly compensation on the multiple sets of aperture detection data based on the multiple sets of aperture detection scene data to obtain multiple aperture detection results;

[0111] A target aperture detection report generation module is used to perform data fusion based on the multiple aperture detection results to generate the target aperture detection report.

[0112] Furthermore, the embodiment of the present application also includes:

[0113] an aperture detection data acquisition module, the aperture detection data acquisition module being configured to traverse the multiple sets of aperture detection data and the multiple sets of aperture detection scene data, extract a first set of aperture detection data, and a first set of aperture detection scene data corresponding to the first set of aperture detection data, wherein the first set of aperture detection scene data includes first aperture detection environment information and first aperture detection device monitoring information;

[0114] an abnormality recognition and evaluation module, configured to perform abnormality recognition and evaluation on the first aperture detection environment information, and determine a first environment abnormality recognition result and a first environment abnormality evaluation coefficient;

[0115] an abnormality identification and evaluation module, the abnormality identification and evaluation module being configured to perform abnormality identification and evaluation on the monitoring information of the first aperture detection device, and obtain a first device abnormality identification result and a first device abnormality evaluation coefficient;

[0116] an environmental anomaly compensation data acquisition module, the environmental anomaly compensation data acquisition module being configured to input the first environmental anomaly recognition result into an environmental anomaly compensation analysis model to obtain first environmental anomaly compensation data if the first environmental anomaly evaluation coefficient is greater than / equal to an environmental anomaly evaluation threshold and the first device anomaly evaluation coefficient is less than the device anomaly evaluation threshold, and to adaptively compensate the first set of aperture detection scene data based on the first environmental anomaly compensation data to obtain a first aperture detection result, and to add the first aperture detection result to the multiple aperture detection results;

[0117] an aperture detection result generating module, configured to input the first device abnormality identification result into a device abnormality compensation analysis model to obtain first device abnormality compensation data if the first environment abnormality evaluation coefficient is less than the environment abnormality evaluation threshold and the first device abnormality evaluation coefficient is greater than or equal to the device abnormality evaluation threshold, and perform adaptive compensation on the first set of aperture detection scene data based on the first device abnormality compensation data to generate the first aperture detection result;

[0118] an adaptive compensation module configured to, if the first environmental anomaly evaluation coefficient is greater than / equal to the environmental anomaly evaluation threshold and the first device anomaly evaluation coefficient is greater than / equal to the device anomaly evaluation threshold, perform anomaly compensation fusion based on the first environmental anomaly identification result and the first device anomaly identification result, according to the environmental anomaly compensation analysis model and the device anomaly compensation analysis model, to obtain first fused anomaly compensation data, and perform adaptive compensation on the first set of aperture detection scene data based on the first fused anomaly compensation data to obtain the first aperture detection result;

[0119] An aperture detection result output module is used to obtain a first detection data mapping instruction if the first environmental abnormality evaluation coefficient is less than the environmental abnormality evaluation threshold, and the first equipment abnormality evaluation coefficient is less than the equipment abnormality evaluation threshold, and output the first set of aperture detection data as the first aperture detection result according to the first detection data mapping instruction.

[0120] For a specific embodiment of a probe card aperture adaptive detection system, please refer to the embodiment of a probe card aperture adaptive detection method described above and will not be repeated here. Each of the above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0121] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor, a memory and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store news data and data such as time attenuation factors. The network interface of the computer device is used to communicate with an external terminal via a network connection. The computer program is executed by the processor to implement an adaptive detection method for the aperture of a probe card.

[0122] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0123] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of a method for adaptively detecting the aperture of a probe card when executing the computer program.

[0124] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the computer program implements the steps of a method for adaptively detecting the aperture of a probe card.

[0125] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0126] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for adaptively detecting the aperture of a probe card, characterized in that: The method comprises: Load the target probe card's specification, model, attribute information, and aperture detection accuracy parameters; Acquire the in-hole image data of the target probe card using an image acquisition device to obtain an in-hole image acquisition result, and perform image enhancement on the in-hole image acquisition result according to a Retinex algorithm to obtain an in-hole image of the target; Based on the probe card in-hole cleaning detector, performing in-hole cleaning verification on the target probe card according to the target in-hole image to obtain a target in-hole cleaning verification result; When the cleaning verification result in the target hole is passed, the target probe card is placed on a stage of a probe card aperture detection device to obtain real-time stage control data and real-time target positioning data; Based on the specification and model attribute information, and in accordance with multi-level detection feature verification constraints, performing detection feature verification on the real-time stage control data and the real-time target positioning data to obtain a detection feature verification result; When the detection feature verification result is qualified, the real-time stage control data and the real-time target positioning data are input into a pre-built detection equipment control decision model to obtain a detection equipment control decision; Based on the adaptive aperture detection algorithm, the target probe card is detected according to the aperture detection accuracy parameter, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report.

2. The method according to claim 1, wherein Based on the probe card in-hole cleaning detector, performing in-hole cleaning verification on the target probe card according to the target in-hole image to obtain the target in-hole cleaning verification result, including: The probe card hole cleaning detector includes a probe card hole cleaning evaluation unit and a probe card hole cleaning verification unit; Inputting the target hole interior image into the probe card hole interior cleaning evaluation unit to obtain the target hole interior cleaning evaluation coefficient; Inputting the target hole cleaning evaluation coefficient into the probe card hole cleaning verification unit to generate the target hole cleaning verification result; The probe card hole cleaning verification unit includes a probe card hole cleaning verification constraint rule, wherein the probe card hole cleaning verification constraint rule includes that when the target hole cleaning evaluation coefficient satisfies the probe card hole cleaning evaluation constraint, the target hole cleaning verification result obtained is passed; The probe card hole cleaning verification constraint rule further includes obtaining a target hole cleaning verification result as a failure when the target hole cleaning evaluation coefficient does not meet the probe card hole cleaning evaluation constraint, and generating a probe card hole cleaning optimization instruction.

3. The method according to claim 2, wherein Inputting the target hole image into the probe card hole cleaning evaluation unit to obtain the target hole cleaning evaluation coefficient includes: Based on big data, load multiple probe card hole cleaning evaluation record data sets; Based on multiple image feature learning operators, the multiple probe card in-hole cleaning evaluation record data sets are trained and tested respectively to obtain multiple in-hole cleaning evaluation models that meet preset in-hole cleaning evaluation accuracy constraints; Obtaining a plurality of in-hole cleaning evaluation accuracy parameters corresponding to the plurality of in-hole cleaning evaluation models, and performing a proportion calculation based on the plurality of in-hole cleaning evaluation accuracy parameters to construct an in-hole cleaning evaluation output incentive model, wherein the in-hole cleaning evaluation output incentive model includes a plurality of in-hole cleaning evaluation output incentive coefficients; embedding the plurality of in-hole cleaning evaluation models and the in-hole cleaning evaluation output excitation model into the in-hole cleaning evaluation unit of the probe card; Inputting the target hole image into the multiple hole cleaning evaluation models to obtain multiple hole cleaning evaluation coefficients; The multiple in-hole cleaning evaluation coefficients are input into the in-hole cleaning evaluation output excitation model, and the multiple in-hole cleaning evaluation coefficients are weighted calculated according to the multiple in-hole cleaning evaluation output excitation coefficients to generate the target in-hole cleaning evaluation coefficient.

4. The method according to claim 1, wherein Based on the specification model attribute information and in accordance with multi-level detection feature verification constraints, detection feature verification is performed on the real-time stage control data and the real-time target positioning data to obtain a detection feature verification result, including: Based on the specification and model attribute information, stage control anomaly evaluation records of similar probe cards are collected to obtain a similar stage control anomaly evaluation record set, and a stage control anomaly evaluation network that satisfies a first convergence constraint is trained based on the similar stage control anomaly evaluation record set; Collecting positioning anomaly evaluation records of the same type of probe cards based on the specification and model attribute information to obtain a similar positioning anomaly evaluation record set, and training a positioning anomaly evaluation network that satisfies a second convergence constraint based on the similar positioning anomaly evaluation record set; Inputting the real-time stage control data into the stage control abnormality evaluation network to obtain a real-time stage control abnormality evaluation coefficient; Inputting the real-time target positioning data into the positioning anomaly evaluation network to obtain a real-time target positioning anomaly evaluation coefficient; Comparing the multi-level detection feature verification constraint with the real-time stage control anomaly evaluation coefficient and the real-time target positioning anomaly evaluation coefficient to generate the detection feature verification result; Wherein, the multi-level detection feature verification constraints include a stage control abnormality evaluation threshold and a target positioning abnormality evaluation threshold; When the real-time stage control abnormality evaluation coefficient is less than the stage control abnormality evaluation threshold, and the real-time target positioning abnormality evaluation coefficient is less than the target positioning abnormality evaluation threshold, the generated detection feature verification result is qualified; When the real-time stage control abnormality evaluation coefficient is greater than / equal to the stage control abnormality evaluation threshold, and / or the real-time target positioning abnormality evaluation coefficient is greater than / equal to the target positioning abnormality evaluation threshold, the generated detection feature verification result is unqualified.

5. The method according to claim 1, wherein The method comprises: When the detection feature verification result is unqualified, generating detection abnormal device and device abnormal feature data; Performing equipment abnormality compensation analysis based on the equipment abnormality feature data to determine equipment abnormality compensation adjustment data; The abnormality detection device is adjusted based on the device abnormality compensation adjustment data.

6. The method according to claim 1, wherein Based on the adaptive aperture detection algorithm, the target probe card is detected according to the aperture detection accuracy parameter, the detection equipment control decision and the probe card aperture detection device to obtain a target aperture detection report, including: Inputting the aperture detection accuracy parameter into a pre-constructed aperture detection frequency measurement table to obtain a matching aperture detection frequency; Taking the matching aperture detection frequency as a detection frequency constraint, controlling the probe card aperture detection device to perform multiple detections on the target probe card according to the detection equipment control decision, to obtain multiple sets of aperture detection data that meet the detection frequency constraint; Collecting real-time environmental information and real-time equipment monitoring information corresponding to the multiple sets of aperture detection data to obtain multiple sets of aperture detection scene data; performing adaptive detection scene anomaly compensation on the multiple sets of aperture detection data based on the multiple sets of aperture detection scene data to obtain multiple aperture detection results; Data fusion is performed based on the multiple aperture detection results to generate the target aperture detection report.

7. The method according to claim 6, wherein Adaptively compensating for abnormalities in detection scenes on the multiple sets of aperture detection data based on the multiple sets of aperture detection scene data to obtain multiple aperture detection results, including: Traversing the multiple sets of aperture detection data and the multiple sets of aperture detection scene data, extracting a first set of aperture detection data and a first set of aperture detection scene data corresponding to the first set of aperture detection data, wherein the first set of aperture detection scene data includes first aperture detection environment information and first aperture detection device monitoring information; Performing anomaly recognition and evaluation on the first aperture detection environment information to determine a first environment anomaly recognition result and a first environment anomaly evaluation coefficient; Performing abnormality identification and evaluation on the monitoring information of the first aperture detection device to obtain a first device abnormality identification result and a first device abnormality evaluation coefficient; If the first environmental anomaly evaluation coefficient is greater than / equal to the environmental anomaly evaluation threshold, and the first device anomaly evaluation coefficient is less than the device anomaly evaluation threshold, inputting the first environmental anomaly recognition result into an environmental anomaly compensation analysis model to obtain first environmental anomaly compensation data, and adaptively compensating the first set of aperture detection scene data based on the first environmental anomaly compensation data to obtain a first aperture detection result, and adding the first aperture detection result to the multiple aperture detection results; If the first environment abnormality evaluation coefficient is less than the environment abnormality evaluation threshold, and the first device abnormality evaluation coefficient is greater than / equal to the device abnormality evaluation threshold, inputting the first device abnormality identification result into the device abnormality compensation analysis model to obtain first device abnormality compensation data, and adaptively compensating the first set of aperture detection scene data based on the first device abnormality compensation data to generate the first aperture detection result; If the first environment abnormality evaluation coefficient is greater than / equal to the environment abnormality evaluation threshold, and the first device abnormality evaluation coefficient is greater than / equal to the device abnormality evaluation threshold, based on the first environment abnormality identification result and the first device abnormality identification result, abnormality compensation fusion is performed according to the environment abnormality compensation analysis model and the device abnormality compensation analysis model to obtain first fused abnormality compensation data, and adaptive compensation is performed on the first set of aperture detection scene data based on the first fused abnormality compensation data to obtain the first aperture detection result; If the first environmental abnormality evaluation coefficient is less than the environmental abnormality evaluation threshold, and the first equipment abnormality evaluation coefficient is less than the equipment abnormality evaluation threshold, a first detection data mapping instruction is obtained, and the first set of aperture detection data is output as the first aperture detection result according to the first detection data mapping instruction.

8. An adaptive detection system for probe card aperture, characterized in that: The system is used to perform the method according to any one of claims 1 to 7, and the system comprises: An attribute information loading module, wherein the attribute information loading module is used to load the specification and model attribute information of the target probe card and the aperture detection accuracy parameters; a target well image acquisition module, configured to acquire in-well image data of the target probe card using an image acquisition device to obtain an in-well image acquisition result, and perform image enhancement on the in-well image acquisition result using a Retinex algorithm to acquire an in-well image of the target; A target hole cleaning verification module, wherein the target hole cleaning verification module is used to perform hole cleaning verification on the target probe card according to the target hole image based on the probe card hole cleaning detector to obtain a target hole cleaning verification result; a control positioning data acquisition module, wherein the control positioning data acquisition module is used to place the target probe card on a stage of a probe card aperture detection device when a cleaning verification result in the target hole is passed, and obtain real-time stage control data and real-time target positioning data; a detection result obtaining module for obtaining the detection result to be verified, the detection result obtaining module being used to perform detection feature verification on the real-time stage control data and the real-time target positioning data based on the specification and model attribute information and in accordance with multi-level detection feature verification constraints to obtain a detection feature verification result; a detection equipment control decision module, wherein when the detection feature verification result is qualified, the detection equipment control decision module is used to input the real-time stage control data and the real-time target positioning data into a pre-built detection equipment control decision model to obtain a detection equipment control decision; A target aperture detection report acquisition module is used to detect the target probe card based on an adaptive aperture detection algorithm, according to the aperture detection accuracy parameters, the detection equipment control decision and the probe card aperture detection device, to obtain a target aperture detection report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images