Intelligent technological process and yield judgment method, system and equipment and medium

Abstract

The invention relates to an intelligent technological process and yield judgment method, system and device and a medium, and the method comprises the steps: obtaining monitoring data of key nodes of a production line, carrying out the preprocessing of the monitoring data, and storing the monitoring data in a historical database; extracting corresponding historical monitoring data from the historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; performing category analysis on the monitoring data to obtain image data and spectral data, and performing visual analysis on the image data to obtain corresponding visual feature data; performing spectral analysis on the spectral data to obtain corresponding hyperspectral data, and correspondingly associating the visual feature data with the hyperspectral data to obtain associated data; and performing comprehensive analysis on the visual feature data and the hyperspectral data according to the associated data to obtain a comprehensive yield. According to the method, comprehensive analysis can be performed by integrating various data sources, so that the accuracy and comprehensiveness of judgment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of process flow, and in particular to an intelligent process flow and yield rate judgment method, system, equipment and medium. Background Art

[0002] In modern manufacturing, improving the yield rate of products is one of the key factors to ensure product quality and production efficiency. Traditional production process monitoring and yield rate judgment methods often rely on manual experience and manual inspection by technicians. This method is not only time-consuming and labor-intensive, but also easily affected by subjective factors, resulting in limited accuracy of the judgment results. With the development of intelligent manufacturing technology, how to use advanced data processing and analysis technology to improve the intelligence level of the process and thus improve the yield rate has become an urgent problem to be solved. The main problems currently exist include: a large amount of monitoring data collected on the production line has not been pre-processed, the data quality is uneven, and it is difficult to use it directly for analysis; there is a lack of effective data correction methods, and historical data cannot be fully utilized to improve the accuracy and reliability of current data; traditional data analysis methods are difficult to process image data and spectral data at the same time, and cannot comprehensively evaluate product quality; existing yield rate judgment methods often rely too much on a single data type and cannot integrate multiple data sources for comprehensive analysis, thereby limiting the accuracy and comprehensiveness of the judgment. Summary of the invention

[0003] The main purpose of the present invention is to provide an intelligent process flow and yield rate judgment method, system, equipment and medium, which can integrate multiple data sources for comprehensive analysis, thereby improving the accuracy and comprehensiveness of the judgment.

[0004] To achieve the above object, the present invention provides an intelligent process flow and yield rate judgment method, comprising: Acquire monitoring data of key nodes of the production line, pre-process the monitoring data, and store them in a historical database; Extracting corresponding historical monitoring data from the historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; Performing category analysis on the monitoring data to obtain image data and spectral data, and performing visual analysis on the image data to obtain corresponding visual feature data; Performing spectral analysis on the spectral data to obtain corresponding hyperspectral data, and correlating the visual feature data with the hyperspectral data to obtain correlated data; The visual feature data and the hyperspectral data are comprehensively analyzed according to the associated data to obtain a comprehensive yield rate.

[0005] Furthermore, the acquisition of monitoring data of key nodes of the production line, preprocessing of the monitoring data, and storage in a historical database includes: Identify key nodes of the production line, determine the locations of multiple key nodes, and deploy monitoring equipment at multiple key node locations; The monitoring equipment is used to collect data at key node locations to obtain original monitoring data, and the original monitoring data is subjected to noise filtering to obtain filtered data; The filtered data is normalized to obtain normalized data; the normalized data is sorted in chronological order to obtain the monitoring data, and the monitoring data is stored in a historical database.

[0006] Further, extracting corresponding historical monitoring data from the historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data, includes: Perform matching analysis on the monitoring data to obtain a node location range, and extract the corresponding historical monitoring data based on the node location range; Performing data standard analysis on the historical monitoring data to obtain a standard reference format, performing anomaly detection on the monitoring data according to the standard reference format, and marking the monitoring data as revised monitoring data when no anomaly is detected; When an abnormality is detected, the abnormal part of the monitoring data is marked, and node identification is performed on the monitoring data marked with the abnormal part to obtain node information, and the corresponding first monitoring data is re-acquired based on the node information. The monitoring data marked with the abnormal part is corrected based on the historical monitoring data and the first monitoring data to obtain the corrected monitoring data.

[0007] Further, the performing category analysis on the monitoring data to obtain image data and spectral data, and performing visual analysis on the image data to obtain corresponding visual feature data, includes: Performing category analysis on the monitoring data according to preset category rules to obtain the image data and the spectral data; Performing image preprocessing on the image data to obtain processed image data; Performing color space conversion and color histogram analysis on the processed image data in sequence to obtain color feature data; Calculate the gray level co-occurrence matrix of the processed image data to obtain the texture feature matrix, calculate the surface roughness and uniformity parameters according to the texture feature matrix, and obtain the texture feature data; Performing edge detection on the processed image data to obtain edge feature data, performing contour extraction based on the edge feature data to obtain contour feature data, and performing shape analysis on the contour feature data to obtain shape parameter data; The color feature data, the texture feature data and the shape parameter data are subjected to feature fusion to obtain fused feature data, and feature dimension reduction processing is performed on the fused feature data to obtain the visual feature data.

[0008] Further, the performing spectral analysis on the spectral data to obtain corresponding hyperspectral data, and correspondingly associating the visual feature data with the hyperspectral data to obtain associated data, comprises: Performing spectral reflectance calculation on the spectral data to obtain spectral reflectance data; The spectral reflectance data is converted into a function curve to obtain a reflectance curve, the first-order derivative of the reflectance curve is calculated, and the reflectance curve is analyzed by continuous wavelet transform to obtain absorption band data, the absorption band data is screened by a threshold method, and the central wavelength, depth and width of each absorption band are recorded, the identified absorption band is compared with the characteristic absorption band of a known substance, and preliminary substance identification is performed to obtain absorption band characteristic data; Integrate and process the spectral reflectance data and the absorption band characteristic data to obtain the hyperspectral data; Performing feature matching on the visual feature data and the hyperspectral data to obtain a feature matching result, and performing spatial registration on the visual feature data and the hyperspectral data according to the feature matching result to obtain registration data; Performing feature fusion on the registration data to obtain fused feature data, performing correlation analysis on the fused feature data to obtain a correlation matrix, and performing feature selection on the fused feature data according to the correlation matrix to obtain preliminary correlation data; The preliminary associated data is subjected to data structuring to obtain structured associated data; the structured associated data is subjected to association rule mining to obtain association rules; and the structured associated data is subjected to feature combination according to the association rules to obtain associated data.

[0009] Furthermore, the visual feature data and the hyperspectral data are comprehensively analyzed based on the associated data to obtain a comprehensive yield rate, including: Performing feature extraction on the associated data to obtain an associated feature vector, and performing dimensionality reduction processing on the associated feature vector to obtain a reduced-dimensional feature vector; A multi-dimensional evaluation matrix is ​​constructed according to the dimension-reduced feature vector, and the multi-dimensional evaluation matrix is ​​regularized to obtain a standardized evaluation matrix; Inputting the visual feature data and the hyperspectral data into a standardized evaluation matrix, and obtaining a preliminary yield rate index through matrix operation; Perform weighted average processing on the preliminary yield rate indicators to obtain the weighted average yield rate, perform exponential smoothing analysis on the historical yield rate data based on the weighted average yield rate, and obtain the time series correction factor; The weighted average yield rate is corrected according to the time series correction factor to obtain the corrected yield rate; Perform quantile analysis on the corrected yield rate to obtain the yield rate distribution range; The corrected yield rate is piecewise linearly interpolated according to the yield rate distribution interval to obtain the final comprehensive yield rate.

[0010] The present invention also provides an intelligent process flow and yield rate judgment system, which is applied to any of the above intelligent process flow and yield rate judgment methods, including: A collection module, which is used to obtain monitoring data of key nodes of the production line, pre-process the monitoring data, and store it in a historical database; An analysis module, the analysis module is used to extract corresponding historical monitoring data from the historical database, and perform data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; An association module, the association module is used to perform category analysis on the monitoring data to obtain image data and spectral data, and perform visual analysis on the image data to obtain corresponding visual feature data; A processing module is used to perform spectral analysis on the spectral data to obtain corresponding hyperspectral data, and to associate the visual feature data with the hyperspectral data to obtain associated data.

[0011] A control module is used to perform a comprehensive analysis on the visual feature data and the hyperspectral data according to the associated data to obtain a comprehensive yield rate.

[0012] The present invention also provides an intelligent process flow and yield rate judgment device, including: Memory, used to store programs; The processor is used to execute the program to implement each step of an intelligent process flow and yield rate judgment method as described in any one of the above items.

[0013] The present invention also provides a storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute any of the above methods.

[0014] The present invention provides the following beneficial effects: By preprocessing the monitoring data of key nodes of the production line, the accuracy and reliability of the data are ensured, laying a solid foundation for subsequent data analysis. By extracting and utilizing historical monitoring data, data correction is performed to improve the accuracy and reliability of current monitoring data, making data analysis more reliable. Visual feature data is obtained by visual analysis of image data, and hyperspectral data is obtained by spectral analysis of spectral data. The two are correlated to achieve a comprehensive evaluation of product quality from different angles. Comprehensive analysis of visual feature data and hyperspectral data can be performed to obtain the comprehensive yield rate, which not only improves the accuracy of judgment, but also integrates multiple data sources to improve the comprehensiveness of judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of an intelligent process flow and yield rate judgment method provided by the present invention; Figure 2 This is a structure diagram of an intelligent process flow and yield rate judgment system provided by the present invention; Figure 3 This is a structure diagram of an intelligent process flow and yield rate judgment equipment provided by the present invention.

[0016] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0019] Reference Figure 1 As shown, the present invention provides an intelligent process flow and yield rate judgment method, including: Step S1: Use sensors, cameras and other equipment to collect monitoring data of key nodes of the production line in real time. The monitoring data includes various types of signals such as images, videos, temperature, and pressure.

[0020] Convert raw data into a format suitable for analysis, such as converting image data into grayscale or RGB images. Label the collected data, such as marking objects in the image, for subsequent analysis. Store the pre-processed monitoring data in a historical database, which can use a relational database (such as MySQL), a NoSQL database (such as MongoDB), or a data warehouse (such as Amazon Redshift).

[0021] Step S2: extracting historical data corresponding to the current monitoring data from the historical database. The historical data is the monitoring records over a period of time in the past, which are used for comparison with the current data.

[0022] Compare current monitoring data with historical monitoring data to find anomalies or deviations in the data. Correct the current monitoring data according to the normal range in historical data, such as through smoothing, missing value filling, etc. Use statistical methods or machine learning algorithms to detect outliers in the data and perform corresponding processing to obtain corrected monitoring data.

[0023] Step S3: dividing the corrected monitoring data into image data and spectral data.

[0024] The image data is preprocessed, such as grayscale, normalization, cropping, etc., to improve the analysis efficiency and accuracy, and key features such as edges, textures, color histograms, etc. are extracted from the image. The most relevant features are selected according to the analysis purpose, and the extracted features are converted into numerical representations. The resulting feature set is called visual feature data.

[0025] Step S4: preprocess the spectral data, such as smoothing, baseline correction, etc., to reduce the influence of noise.

[0026] Extract key features from spectral data, such as absorption peak position, intensity, etc.

[0027] The feature set obtained through spectral analysis is called hyperspectral data, which includes information such as the chemical composition and physical properties of the product.

[0028] The visual feature data is associated with the hyperspectral data, which is usually achieved by establishing a common index or label to ensure that each set of feature data can accurately correspond to the same product instance.

[0029] The visual feature data and the hyperspectral data are associated to form associated data.

[0030] Step S5: Comprehensively analyze the visual feature data and the hyperspectral data based on the associated data to obtain a comprehensive yield rate.

[0031] The intelligent process flow and yield rate judgment method provided by the present invention ensures the accuracy and reliability of the data by preprocessing the monitoring data of the key nodes of the production line, laying a solid foundation for subsequent data analysis. By extracting and utilizing the historical monitoring data, data correction is performed, the accuracy and reliability of the current monitoring data are improved, and the data analysis is made more reliable. By performing visual analysis on the image data to obtain visual feature data, performing spectral analysis on the spectral data to obtain hyperspectral data, and correlating the two, a comprehensive evaluation of product quality from different angles is achieved. By performing a comprehensive analysis on the visual feature data and the hyperspectral data, the comprehensive yield rate is obtained, thereby improving the accuracy of the judgment, and can also integrate multiple data sources to improve the comprehensiveness of the judgment.

[0032] In one embodiment, the monitoring data of key nodes of the production line are obtained, the monitoring data are preprocessed, and stored in a historical database, including: Identify key nodes of the production line, determine the locations of multiple key nodes, and deploy monitoring equipment at multiple key node locations. The monitoring equipment includes cameras and sensors, etc., which are used to collect data on the production line in real time.

[0033] The monitoring equipment is used to collect data at key node locations to obtain original monitoring data, and the original monitoring data is subjected to noise filtering to obtain filtered data; Normalize the filtered data to obtain normalized data; sort the normalized data in chronological order to obtain monitoring data, and store the monitoring data in a historical database. The historical database can be a relational database (such as MySQL), a NoSQL database (such as MongoDB), or a time series database (such as InfluxDB), etc. It can be indexed according to timestamps during storage to facilitate subsequent query and analysis.

[0034] This embodiment can effectively remove noise interference in the data and improve the purity of the data by performing noise filtering on the original monitoring data. Normalizing the data to the same numerical range helps to eliminate the impact of dimensional differences and improve the accuracy and efficiency of data analysis. Sorting the data in chronological order ensures the time series of the data, which is conducive to subsequent time series analysis and prediction.

[0035] All monitoring data are pre-processed in a unified way to ensure data consistency and comparability. The processed monitoring data is stored in the historical database to facilitate subsequent query, analysis and mining. Through the database indexing mechanism, the required data can be quickly accessed and retrieved, improving data processing efficiency.

[0036] In one embodiment, extracting corresponding historical monitoring data from a historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data includes: Perform matching analysis on the current monitoring data to determine the location range of key nodes of the production line corresponding to the data, and extract the corresponding historical monitoring data from the historical database based on the determined node location range.

[0037] Perform data standard analysis on historical monitoring data to obtain a standard reference format.

[0038] Anomaly detection is performed on the current monitoring data according to the standard reference format to check whether the data has anomalies. If no anomalies are detected, the correction step is skipped and the monitoring data is marked as corrected monitoring data.

[0039] When an anomaly is detected, the anomaly of the monitoring data is marked, and node identification is performed on the monitoring data marked with the anomaly to obtain node information.

[0040] The corresponding first monitoring data is reacquired according to the node information, and the monitoring data marked with abnormalities is corrected based on the historical monitoring data and the first monitoring data to obtain corrected monitoring data.

[0041] This embodiment can detect abnormalities in the data in a timely manner and ensure the accuracy and reliability of the data by performing abnormality detection on the monitoring data. By correcting the monitoring data based on historical monitoring data, abnormal data can be eliminated or corrected to improve data quality. By performing data standard analysis on historical monitoring data, a standard reference format is obtained to ensure the consistency and comparability of the data. The corrected monitoring data is consistent with the historical data, which helps to make more accurate comparisons and analyses. By processing the monitoring data of key nodes, important links on the production line can be monitored more accurately and potential problems can be discovered in a timely manner. By detecting and correcting abnormalities in the monitoring data, abnormal situations on the production line can be quickly identified, and timely measures can be taken to prevent the occurrence of failures.

[0042] In one embodiment, the monitoring data is subjected to category analysis to obtain image data and spectral data, and the image data is subjected to visual analysis to obtain corresponding visual feature data, including: Perform category analysis on monitoring data through preset type rules to obtain image data and spectral data; Performing image preprocessing on the image data to obtain processed image data; Performing color space conversion and color histogram analysis on the processed image data in sequence to obtain color feature data; Calculate the gray level co-occurrence matrix of the processed image data to obtain the texture feature matrix, calculate the surface roughness and uniformity parameters according to the texture feature matrix, and obtain the texture feature data; Performing edge detection on the processed image data to obtain edge feature data, performing contour extraction based on the edge feature data to obtain contour feature data, and performing shape analysis on the contour feature data to obtain shape parameter data; The color feature data, texture feature data and shape parameter data are subjected to feature fusion to obtain fused feature data, and feature dimension reduction processing is performed on the fused feature data to obtain visual feature data.

[0043] This embodiment can process different types of data separately by performing category analysis on monitoring data, ensuring the pertinence and effectiveness of subsequent processing. By preprocessing image data, noise can be removed, image quality can be improved, and a good foundation can be laid for subsequent analysis. Color features can be extracted through color space conversion and color histogram analysis, which is very useful for identifying color differences between different materials or products. Texture information on the surface of an object can be captured through grayscale co-occurrence matrix calculation and surface roughness / uniformity parameter calculation, which is particularly useful for detecting surface defects or distinguishing products of different materials. Shape features of an object can be extracted through edge detection, contour extraction, and shape analysis, which is very important for identifying the shape and size of an object. Fusion of color feature data, texture feature data, and shape parameter data can obtain a more comprehensive feature description, which helps to improve the accuracy of recognition and classification.

[0044] In one embodiment, spectral analysis is performed on the spectral data to obtain corresponding hyperspectral data, and visual feature data is correspondingly associated with the hyperspectral data to obtain associated data, including: Calculating the spectral reflectance of the spectral data to obtain spectral reflectance data; The spectral reflectance data is converted into a function curve to obtain a reflectance curve. The first-order derivative of the reflectance curve is calculated and the reflectance curve is analyzed using a continuous wavelet transform to obtain absorption band data. The absorption band data is screened using a threshold method, and the central wavelength, depth and width of each absorption band are recorded. The identified absorption band is compared with the characteristic absorption band of a known substance to perform preliminary substance identification and obtain absorption band characteristic data. The spectral reflectance data and the absorption band characteristic data are integrated and processed to obtain the hyperspectral data; Performing feature matching on the visual feature data and the hyperspectral data to obtain a feature matching result, and performing spatial registration on the visual feature data and the hyperspectral data according to the feature matching result to obtain registration data; Perform feature fusion on the registration data to obtain fused feature data, perform correlation analysis on the fused feature data to obtain a correlation matrix, perform feature selection on the fused feature data based on the correlation matrix to obtain preliminary correlation data; The preliminary associated data is subjected to data structuring to obtain structured associated data; the structured associated data is subjected to association rule mining to obtain association rules; and the structured associated data is subjected to feature combination according to the association rules to obtain associated data.

[0045] This embodiment improves the accuracy and reliability of data by calculating the spectral reflectance of spectral data, ensuring the accuracy of subsequent analysis. By performing first-order derivative calculation and continuous wavelet transform analysis on the reflectance curve, the absorption band characteristics are effectively extracted, and the accuracy of material identification is improved. By comparing the identified absorption band with the characteristic absorption band of known substances, preliminary material identification is achieved, which is crucial for quality control on the production line. By recording information such as the center wavelength, depth and width of the absorption band, detailed absorption band characteristic data is constructed, providing rich information for subsequent analysis. The spectral reflectance data and the absorption band characteristic data are integrated and processed to obtain hyperspectral data, which improves the integrity and availability of the data. By performing feature matching and spatial registration on visual feature data and hyperspectral data, the corresponding relationship between the two is ensured, and the consistency and comparability of the data are improved. By fusing visual feature data and hyperspectral data, more comprehensive fusion feature data is obtained, which improves the accuracy of recognition and classification.

[0046] In one embodiment, the visual feature data and the hyperspectral data are comprehensively analyzed based on the correlation data to obtain a comprehensive yield rate, including: Perform feature extraction on the associated data to obtain the associated feature vector, and perform dimensionality reduction processing on the associated feature vector. Methods such as principal component analysis (PCA) and linear discriminant analysis (LDA) can be used to obtain the reduced dimensionality feature vector.

[0047] A multidimensional evaluation matrix is ​​constructed based on the reduced-dimensional feature vector. The multidimensional evaluation matrix includes the importance of different features. The multidimensional evaluation matrix is ​​regularized to obtain a standardized evaluation matrix.

[0048] The visual feature data and hyperspectral data are input into the standardized evaluation matrix, and the preliminary yield index is obtained through matrix operation.

[0049] The preliminary yield rate index is weighted averaged to obtain the weighted average yield rate. The historical yield rate data is subjected to exponential smoothing analysis based on the weighted average yield rate to obtain the time series correction factor.

[0050] The weighted average yield rate is corrected according to the time series correction factor to obtain the corrected yield rate.

[0051] Perform quantile analysis on the corrected yield rate to obtain the yield rate distribution range.

[0052] The corrected yield rate is piecewise linearly interpolated according to the yield rate distribution interval to obtain the final comprehensive yield rate.

[0053] This embodiment improves data processing efficiency and retains key information by performing feature extraction and dimensionality reduction on associated data; constructs a multi-dimensional evaluation matrix and ensures the comparability and consistency between different features through regularization processing, thereby improving the accuracy of evaluation; performs weighted average processing and exponential smoothing analysis on the preliminary yield rate indicator, takes into account the changing trend of historical data, and improves the accuracy of yield rate prediction; obtains the final comprehensive yield rate by performing quantile analysis and piecewise linear interpolation on the corrected yield rate, realizes refined management and optimization of the production process, helps to improve production efficiency and product quality, reduce costs, achieve optimal allocation of resources, and provide a scientific decision-making basis for management.

[0054] The present invention also provides an intelligent process flow and yield rate judgment system, which is applied to any of the above intelligent process flow and yield rate judgment methods, including: The acquisition module is used to obtain monitoring data of key nodes of the production line, pre-process the monitoring data, and store it in the historical database; An analysis module, the analysis module is used to extract corresponding historical monitoring data from a historical database, and perform data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; The association module is used to perform category analysis on the monitoring data to obtain image data and spectral data, and to perform visual analysis on the image data to obtain corresponding visual feature data; The processing module is used to perform spectral analysis on the spectral data to obtain corresponding hyperspectral data, and to associate the visual feature data with the hyperspectral data to obtain associated data.

[0055] The control module is used to perform comprehensive analysis on the visual feature data and the hyperspectral data based on the associated data to obtain a comprehensive yield rate.

[0056] The present invention provides an intelligent process flow and yield rate judgment system, which ensures the accuracy and reliability of the data by preprocessing the monitoring data of the key nodes of the production line, and lays a solid foundation for subsequent data analysis. By extracting and utilizing historical monitoring data, data correction is performed, the accuracy and reliability of the current monitoring data are improved, and the data analysis is made more reliable. By performing visual analysis on image data to obtain visual feature data, performing spectral analysis on spectral data to obtain hyperspectral data, and correlating the two, a comprehensive evaluation of product quality from different angles is achieved. By performing comprehensive analysis on visual feature data and hyperspectral data, a comprehensive yield rate is obtained, thereby improving the accuracy of judgment, and multiple data sources can be integrated to improve the comprehensiveness of judgment.

[0057] The present invention also provides an intelligent process flow and yield rate judgment device, including: Memory, used to store programs; The processor is used to execute the program to implement each step of an intelligent process flow and yield rate judgment method as described in any one of the above items.

[0058] In this embodiment, the processor and the memory may be connected via a bus or other means. The memory may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a read-only memory, a flash memory, a hard disk, or a solid-state drive. The processor may be a general-purpose processor, such as a central processing unit, a digital signal processor, an application-specific integrated circuit, or one or more integrated circuits configured to implement the embodiments of the present invention.

[0059] The present invention also provides a storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute any of the above methods.

[0060] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the system and each module described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0061] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent process flow and yield rate judgment method, characterized in that: include: Acquire monitoring data of key nodes of the production line, pre-process the monitoring data, and store them in a historical database; Extracting corresponding historical monitoring data from the historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; Performing category analysis on the monitoring data to obtain image data and spectral data, and performing visual analysis on the image data to obtain corresponding visual feature data; Performing spectral analysis on the spectral data to obtain corresponding hyperspectral data, and correlating the visual feature data with the hyperspectral data to obtain correlated data; The visual feature data and the hyperspectral data are comprehensively analyzed according to the associated data to obtain a comprehensive yield rate.

2. The intelligent process flow and yield rate judgment method according to claim 1 is characterized in that: The step of obtaining monitoring data of key nodes of the production line, preprocessing the monitoring data, and storing the monitoring data in a historical database includes: Identify key nodes of the production line, determine the locations of multiple key nodes, and deploy monitoring equipment at multiple key node locations; The monitoring equipment is used to collect data at key node locations to obtain original monitoring data, and the original monitoring data is subjected to noise filtering to obtain filtered data; The filtered data is normalized to obtain normalized data; the normalized data is sorted in chronological order to obtain the monitoring data, and the monitoring data is stored in a historical database.

3. The intelligent process flow and yield rate judgment method according to claim 1 is characterized in that: The extracting corresponding historical monitoring data from the historical database, and performing data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data, includes: Perform matching analysis on the monitoring data to obtain a node location range, and extract the corresponding historical monitoring data based on the node location range; Performing data standard analysis on the historical monitoring data to obtain a standard reference format, performing anomaly detection on the monitoring data according to the standard reference format, and marking the monitoring data as revised monitoring data when no anomaly is detected; When an abnormality is detected, the abnormal part of the monitoring data is marked, and node identification is performed on the monitoring data marked with the abnormal part to obtain node information, and the corresponding first monitoring data is re-acquired based on the node information. The monitoring data marked with the abnormal part is corrected based on the historical monitoring data and the first monitoring data to obtain the corrected monitoring data.

4. The intelligent process flow and yield rate judgment method according to claim 1 is characterized in that: The performing category analysis on the monitoring data to obtain image data and spectral data, and performing visual analysis on the image data to obtain corresponding visual feature data, includes: Performing category analysis on the monitoring data according to preset category rules to obtain the image data and the spectral data; Performing image preprocessing on the image data to obtain processed image data; Performing color space conversion and color histogram analysis on the processed image data in sequence to obtain color feature data; Calculate the gray level co-occurrence matrix of the processed image data to obtain the texture feature matrix, calculate the surface roughness and uniformity parameters according to the texture feature matrix, and obtain the texture feature data; Performing edge detection on the processed image data to obtain edge feature data, performing contour extraction based on the edge feature data to obtain contour feature data, and performing shape analysis on the contour feature data to obtain shape parameter data; The color feature data, the texture feature data and the shape parameter data are subjected to feature fusion to obtain fused feature data, and feature dimension reduction processing is performed on the fused feature data to obtain the visual feature data.

5. The intelligent process flow and yield rate judgment method according to claim 4 is characterized in that: The performing spectral analysis on the spectral data to obtain corresponding hyperspectral data, and correlating the visual feature data with the hyperspectral data to obtain associated data, comprises: Performing spectral reflectance calculation on the spectral data to obtain spectral reflectance data; The spectral reflectance data is converted into a function curve to obtain a reflectance curve, the first-order derivative of the reflectance curve is calculated, and the reflectance curve is analyzed by continuous wavelet transform to obtain absorption band data, the absorption band data is screened by a threshold method, and the central wavelength, depth and width of each absorption band are recorded, the identified absorption band is compared with the characteristic absorption band of a known substance, and preliminary substance identification is performed to obtain absorption band characteristic data; Integrate and process the spectral reflectance data and the absorption band characteristic data to obtain the hyperspectral data; Performing feature matching on the visual feature data and the hyperspectral data to obtain a feature matching result, and performing spatial registration on the visual feature data and the hyperspectral data according to the feature matching result to obtain registration data; Performing feature fusion on the registration data to obtain fused feature data, performing correlation analysis on the fused feature data to obtain a correlation matrix, and performing feature selection on the fused feature data according to the correlation matrix to obtain preliminary correlation data; The preliminary associated data is subjected to data structuring to obtain structured associated data; the structured associated data is subjected to association rule mining to obtain association rules; and the structured associated data is subjected to feature combination according to the association rules to obtain associated data.

6. The intelligent process flow and yield rate judgment method according to claim 1, characterized in that: The step of comprehensively analyzing the visual feature data and the hyperspectral data based on the associated data to obtain a comprehensive yield rate includes: Performing feature extraction on the associated data to obtain an associated feature vector, and performing dimensionality reduction processing on the associated feature vector to obtain a reduced-dimensional feature vector; A multi-dimensional evaluation matrix is ​​constructed according to the dimension-reduced feature vector, and the multi-dimensional evaluation matrix is ​​regularized to obtain a standardized evaluation matrix; Inputting the visual feature data and the hyperspectral data into a standardized evaluation matrix, and obtaining a preliminary yield rate index through matrix operation; Perform weighted average processing on the preliminary yield rate indicators to obtain the weighted average yield rate, perform exponential smoothing analysis on the historical yield rate data based on the weighted average yield rate, and obtain the time series correction factor; The weighted average yield rate is corrected according to the time series correction factor to obtain the corrected yield rate; Perform quantile analysis on the corrected yield rate to obtain the yield rate distribution range; The corrected yield rate is piecewise linearly interpolated according to the yield rate distribution interval to obtain the final comprehensive yield rate.

7. An intelligent process flow and yield rate judgment system, applied to the intelligent process flow and yield rate judgment method of any one of claims 1 to 6 above, characterized in that: include: A collection module, which is used to obtain monitoring data of key nodes of the production line, pre-process the monitoring data, and store it in a historical database; An analysis module, the analysis module is used to extract corresponding historical monitoring data from the historical database, and perform data correction on the monitoring data based on the historical monitoring data to obtain corrected monitoring data; An association module, the association module is used to perform category analysis on the monitoring data to obtain image data and spectral data, and perform visual analysis on the image data to obtain corresponding visual feature data; A processing module, the processing module is used to perform spectral analysis on the spectral data to obtain corresponding hyperspectral data, and to associate the visual feature data with the hyperspectral data to obtain associated data; A control module is used to perform a comprehensive analysis on the visual feature data and the hyperspectral data according to the associated data to obtain a comprehensive yield rate.

8. An intelligent process flow and yield rate judgment device, characterized in that: include: Memory, used to store programs; A processor is used to execute the program to implement each step of an intelligent process flow and yield rate judgment method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: Computer instructions are stored, and the computer instructions are used to make a computer execute the method according to any one of claims 1 to 6.

Images