A data management method and system for an intelligent plate processing factory

By collecting image and sensor data from different points of interest, and performing information fusion and anomaly detection, the problem of misjudgment caused by single-angle data management in the sheet metal processing process is solved, achieving more accurate quality control and anomaly monitoring, and improving production efficiency and product quality.

CN118628988BActive Publication Date: 2024-11-15GUANGDONG XG INTELLIGENT SYST CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411106794.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-15
Estimated Expiration
2044-08-13

AI Technical Summary

Technical Problem

In existing technologies, data management and analysis in the sheet metal processing process are often conducted from a single perspective, leading to a high risk of misjudgment and the inability to detect potential problems in a timely and accurate manner.

Method used

Image and sensor data are collected at different points of interest during the sheet material processing. Information fusion and anomaly detection are performed. Abnormal areas are identified by information entropy and binarization processing, and comprehensive analysis is conducted in conjunction with sensor data.

Benefits of technology

It enables comprehensive monitoring of the board processing process, reduces misjudgments, improves the accuracy and timeliness of anomaly detection, and enhances product quality and production efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118628988B_ABST
    Figure CN118628988B_ABST
Patent Text Reader

Abstract

The embodiment of the present invention provides a data management method and system for an intelligent sheet metal processing factory, which belongs to the field of data processing technology. The method includes: collecting a first image corresponding to a target object at a first focus point and a second image corresponding to a second focus point during sheet metal processing, and simultaneously obtaining initial sensor data corresponding to the target object under a target sensor; performing information fusion on the first image and the second image to obtain a target image; performing anomaly detection on the target image using information entropy and binarization processing to obtain a target area corresponding to the target image; performing anomaly detection on the initial sensor data to obtain target sensor data; performing information fusion on the target area and the target sensor data to obtain a data management result corresponding to the target object. This solves the problem that the prior art only manages and analyzes the data of the sheet metal processing process from a single perspective, which greatly increases the risk of misjudgment and causes potential problems to be unable to be discovered in a timely and accurate manner.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data management method and system for an intelligent plate processing factory. Background Art

[0002] In the sheet metal processing process, the application of intelligent monitoring technology has significantly improved product quality, reduced production costs and risks, and enhanced production efficiency and customer satisfaction. However, existing technologies often only manage and analyze data from a single perspective, which greatly increases the risk of misjudgment and prevents potential problems from being discovered in a timely and accurate manner. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide a data management method and system for an intelligent sheet metal processing factory, aiming to solve the problem that the data of the sheet metal processing process is often managed and analyzed only from a single perspective in related technologies, which greatly increases the risk of misjudgment and causes potential problems to fail to be discovered in a timely and accurate manner.

[0004] In a first aspect, an embodiment of the present invention provides a data management method for an intelligent plate processing factory, comprising:

[0005] Collecting a first image corresponding to a first focus point and a second image corresponding to a second focus point of a target object during plate processing, and simultaneously obtaining initial sensor data corresponding to the target object under a target sensor;

[0006] Performing information fusion on the first image and the second image to obtain a target image;

[0007] Performing abnormality detection on the target image by using information entropy and binarization processing to obtain a target area corresponding to the target image;

[0008] Performing outlier detection on the initial sensor data to obtain target sensor data;

[0009] The target area and the target sensor data are subjected to information fusion to obtain the data management result corresponding to the target object.

[0010] In a second aspect, an embodiment of the present invention provides a data management system for an intelligent plate processing factory, including:

[0011] A data acquisition module, used to acquire a first image corresponding to a first focus point and a second image corresponding to a second focus point of a target object during plate processing, and simultaneously obtain initial sensor data corresponding to the target object under a target sensor;

[0012] An information fusion module, used for performing information fusion on the first image and the second image to obtain a target image;

[0013] A region segmentation module is used to perform abnormality detection on the target image by using information entropy and binarization processing to obtain a target region corresponding to the target image;

[0014] An abnormality monitoring module, used for performing abnormal value detection on the initial sensor data to obtain target sensor data;

[0015] A result determination module is used to fuse the target area and the target sensor data to obtain the data management result corresponding to the target object.

[0016] The embodiment of the present invention provides a data management method and system for an intelligent plate processing factory, the method comprising acquiring images at different points of interest and obtaining sensor data at the same time, so as to obtain information of a target object from multiple angles. This all-round data collection ensures comprehensive monitoring of the target object and reduces the problem that may be missed by a single perspective. Then, the first image and the second image are subjected to information fusion to generate a target image. This fusion can integrate the details of different points of interest, provide a more comprehensive view of the target object, and help to more accurately identify and analyze potential problem areas. By applying information entropy and binarization processing to the target image, anomaly detection is performed. This process helps to accurately identify abnormal areas in the target image. Information entropy analysis provides a measure of the amount of information in the image, while binarization processing simplifies the identification of abnormal areas and effectively improves the accuracy of anomaly detection. Then, anomaly detection is performed on the initial sensor data, and the abnormality of the target object under sensor acquisition can be found, thereby finding the abnormality of the target object under sensor acquisition, and finally the target area is fused with the target sensor data to obtain the data management result of the target object. This comprehensive analysis can more accurately evaluate the state of the target object, thereby providing more comprehensive quality control and anomaly monitoring. This can effectively improve the monitoring and anomaly detection capabilities of target objects in the sheet metal processing process, ensure timely discovery and accurate analysis of problems, and thus improve product quality, reduce production costs, and optimize production efficiency. It also solves the problem that related technologies often manage and analyze sheet metal processing data from only a single perspective, which greatly increases the risk of misjudgment and causes potential problems to be unable to be discovered in a timely and accurate manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A schematic diagram of a data management method for an intelligent plate processing factory provided by an embodiment of the present invention;

[0019] Figure 2 A schematic diagram of the module structure of a data management system for an intelligent plate processing factory provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0022] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0023] The embodiment of the present invention provides a data management method and system for a smart sheet metal processing factory. The data management method for a smart sheet metal processing factory can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0024] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0025] Please refer to Figure 1 , Figure 1 A schematic flow chart of a data management method for an intelligent plate processing factory provided in an embodiment of the present invention.

[0026] like Figure 1 As shown, the data management method of the intelligent plate processing factory includes steps S101 to S105.

[0027] Step S101, collecting a first image corresponding to a target object at a first focus point and a second image corresponding to a second focus point during plate processing, and simultaneously obtaining initial sensor data corresponding to the target object under a target sensor.

[0028] Exemplarily, the target object is a physical plate obtained by plate processing. Since the target object is relatively large, when directly using a camera device to collect information, abnormal information in the target object may not be captured or the abnormal information may be blurred, affecting subsequent judgment. Therefore, in the present application, the target object is divided into two, and two focus points are set, that is, the focus points when photographing the target object, so as to obtain a first image corresponding to the target object at the first focus point and a second image corresponding to the second focus point during the plate processing process, wherein the first image and the second image can both collect all the information of the target image, but the focus points when the first image and the second image are collected are different.

[0029] For example, when the material or thickness of the target object obtained after the plate is processed is different, the initial sensor data obtained when data is collected using a temperature sensor, a pressure sensor or other sensor related to the processing process has different properties. Therefore, while obtaining the first image and the second image, the application also obtains the initial sensor data corresponding to the target object under the target sensor, thereby providing support for subsequent monitoring of whether there is an abnormality in the target object.

[0030] For example, target sensors are used to collect panel data (size, processing technology), equipment operation (start and stop time, operating hours, temperature, etc.), status data, and key workstation video monitoring data in the workshop, so as to obtain the corresponding initial sensor data under the target sensor through the interface.

[0031] Step S102: performing information fusion on the first image and the second image to obtain a target image.

[0032] Exemplarily, a feature matching algorithm (such as SIFT, SURF) or an image-based geometric transformation (such as affine transformation, perspective transformation) is used to align the first image and the second image, and then the information of the first image and the second image is fused into a target image by weighted averaging.

[0033] In some embodiments, the information fusion of the first image and the second image to obtain the target image includes: extracting texture features of the first image to obtain a first texture image and extracting texture features of the second image to obtain a second texture image; determining a first target window to obtain a first local image corresponding to the first image in the first target window and a second local image corresponding to the first texture image in the first target window; determining a first pixel mean corresponding to a first pixel value in the first local image and determining a second pixel mean and a first pixel variance corresponding to a second pixel value in the second local image; determining a first blurriness, and establishing a first pixel association relationship between the first image and the first texture image corresponding to the first target window according to the first blurriness, the first pixel mean, the second pixel mean and the first pixel variance; determining a first pixel association relationship between the first image and the first texture image according to the first image The first image is pixel-enhanced according to the pixel association relationship to obtain a first enhanced image; a second target window is determined to obtain a third local image corresponding to the second image in the second target window and a fourth local image corresponding to the second texture image in the second target window; a third pixel mean corresponding to the third pixel value in the third local image is determined and a fourth pixel mean and a second pixel variance corresponding to the fourth pixel value in the fourth local image are determined; a second blurriness is determined, and a second pixel association relationship between the second image and the second texture image corresponding to the second target window is established according to the second blurriness, the third pixel mean, the fourth pixel mean and the second pixel variance; the second image is pixel-enhanced according to the second pixel association relationship to obtain a second enhanced image; information fusion is performed according to the first enhanced image and the second enhanced image to obtain the target image.

[0034] Exemplarily, a texture feature extraction algorithm (such as gray level co-occurrence matrix, local binary pattern, etc.) is used to extract texture features from the first image to obtain a first texture image, and the same texture feature extraction algorithm is applied to the second image to obtain a second texture image.

[0035] Exemplarily, a first target window is defined in the first image and the first texture image, and then an area corresponding to the first target window is extracted from the first image to be determined as a first partial image, and an area corresponding to the first target window is extracted from the first texture image to be determined as a second partial image.

[0036] Exemplarily, the sum and average of each first pixel value in the first local image are calculated to obtain a first pixel mean, and the sum and average of each second pixel value in the second local image are calculated to obtain a second pixel mean, and the first pixel variance corresponding to the second local image is determined based on the second pixel mean.

[0037] Exemplarily, the first blurriness is calculated by analyzing the texture features of the first partial image and the second partial image. Thus, the first coefficient and the second coefficient in the first pixel association relationship are determined according to the first blurriness, the first pixel mean, the second pixel mean, the first pixel variance, and the first pixel value and the second pixel value, thereby determining the first pixel association relationship according to the first coefficient and the second coefficient. For example, the first coefficient and the second coefficient are obtained according to the following formula:

[0038]

[0039]

[0040] in, represents the first coefficient in the first pixel association relationship, represents the total number of pixels in the first target window, represents the first pixel value corresponding to the i-th position in the first local image, represents the second pixel value corresponding to the i-th position in the second local image, represents the first pixel mean and represents the second pixel mean and represents the first pixel variance, represents the first fuzziness, Represents the second coefficient in the first pixel association relationship.

[0041] Exemplarily, the first local image in the first image is pixel enhanced according to the first pixel association relationship, and so on, the first target window is slid in the first image to obtain the first pixel association relationship corresponding to each first target window, thereby obtaining the first enhanced image corresponding to all the first images.

[0042] Exemplarily, a second target window is defined in the second image and the second texture image, and then an area corresponding to the second target window is extracted from the second image to be determined as a third local image, and an area corresponding to the second target window is extracted from the second texture image to be determined as a fourth local image.

[0043] Exemplarily, the sum and average of each third pixel value in the third local image are calculated to obtain a third pixel mean, and the sum and average of each fourth pixel value in the fourth local image are calculated to obtain a fourth pixel mean, and the second pixel variance corresponding to the fourth local image is determined based on the fourth pixel mean.

[0044] Exemplarily, the second blur is calculated by analyzing the texture features of the third partial image and the fourth partial image. Thus, the third coefficient and the fourth coefficient in the second pixel association relationship are determined according to the second blur, the third pixel mean, the fourth pixel mean, and the second pixel variance, and the third pixel value and the fourth pixel value, and the second pixel association relationship is determined according to the third coefficient and the fourth coefficient. For example, the third coefficient and the fourth coefficient are obtained according to the following formula:

[0045]

[0046]

[0047] in, represents the third coefficient in the second pixel association relationship, T represents the total number of pixels in the second target window, represents the third pixel value corresponding to the i-th position in the third local image, represents the fourth pixel value corresponding to the i-th position in the fourth partial image, represents the third pixel mean and represents the fourth pixel mean and u represents the second pixel variance, represents the second ambiguity, Represents the fourth coefficient in the second pixel association relationship.

[0048] Exemplarily, the third local image in the third image is pixel enhanced according to the second pixel association relationship, and so on, the second target window is slid in the third image to obtain the second pixel association relationship corresponding to each second target window, thereby obtaining the second enhanced image corresponding to all the second images.

[0049] Exemplarily, the first enhanced image and the second enhanced image are fused by weighted averaging to generate a target image.

[0050] Specifically, texture feature extraction helps capture detailed information in images, such as texture patterns and structures. This helps to improve the level of detail of the image, making the texture information in the local area clearer. By establishing pixel associations and performing pixel enhancement, the blur and lack of contrast in the image can be corrected, thereby improving the visual effect and information readability of the image. By fusing the enhancement effects of the two images, the useful information in the two images can be combined to reduce the defects that may exist in a single image. This can obtain a more comprehensive and clearer target image. This provides support for the subsequent realization of abnormal detection of the target object.

[0051] In some embodiments, the information fusion according to the first enhanced image and the second enhanced image to obtain the target image includes: comparing the first enhanced image and the first image pixel by pixel according to a first bitwise comparison algorithm to obtain a first comparison map; comparing the second enhanced image and the second image pixel by pixel according to a second bitwise comparison algorithm to obtain a second comparison map, wherein the comparison methods of the first bitwise comparison algorithm and the second bitwise comparison algorithm are opposite; fusing the first comparison map and the second comparison map and performing denoising processing to obtain a target comparison map; fusing the first image and the second image according to the target comparison map to obtain the target image.

[0052] Exemplarily, a bitwise comparison algorithm (e.g., AND, OR, XOR, etc.) is selected to compare the pixel differences between the first enhanced image and the first image, and then the first bitwise comparison algorithm is applied to compare the corresponding pixels of the first enhanced image and the first image to highlight the differences between the two images, obtain the bright and dark areas in the first image, and thus generate a first comparison map, wherein the first comparison map is used to represent areas with more information content in the first image, thereby providing support for subsequent image fusion.

[0053] Exemplarily, a different bitwise comparison algorithm is selected (for example, if the first bitwise comparison algorithm is AND, the second bitwise comparison algorithm may be OR or XOR) to compare the pixel differences between the second enhanced image and the second image, and then the second bitwise comparison algorithm is applied to compare the corresponding pixels of the second enhanced image and the second image to highlight the differences in the images and obtain the dark and light areas in the second image, thereby generating a second comparison map, wherein the second comparison map is used to represent areas with more information content in the second image, thereby providing support for subsequent image fusion.

[0054] Exemplarily, the pixels of the brightness area in the first comparison image in the first image and the pixels of the brightness area in the second comparison image in the second image are merged according to the pixel positions to obtain a fused target image, that is, the fused target image contains the best features and information of the first image and the second image, so that the fused target image has all the best information under the first focus point and the second focus point, and also makes the target object have more accurate image information.

[0055] Specifically, by using a bitwise comparison algorithm, the differences between images can be clearly detected and highlighted. This helps to identify subtle changes or features in the image, and using the opposite bitwise comparison algorithm can analyze image differences from different angles, helping to more fully understand the differences in the image. Thus, by generating a first comparison image and a second comparison image, difference information under different viewing angles can be obtained. This method helps to improve the contrast, making potential changes and details in the image more obvious, thereby providing richer information for subsequent analysis. Finally, by using the first comparison image and the second comparison image to fuse the first image and the second image for image fusion, the features of the first image and the second image can be more accurately synthesized, thereby obtaining a target image so that the target object has more accurate image information.

[0056] Step S103: perform abnormality detection on the target image by using information entropy and binarization processing to obtain a target area corresponding to the target image.

[0057] Exemplarily, if the target image is in color, the information entropy under different channels is calculated, and then the information entropy under different channels is averaged to obtain an average entropy value, and then the target image is binarized according to the average entropy value and a preset value, so as to obtain the corresponding abnormal area in the binarized image, and then the information of the corresponding position of the abnormal area in the target image is determined as the target area.

[0058] In some embodiments, the use of information entropy and binarization processing to perform anomaly detection on the target image to obtain a target area corresponding to the target image includes: determining a target radius corresponding to an area circle and a target number corresponding to the area circle; obtaining a related pixel position associated with each target pixel position in the target image according to the target radius and the target number; determining the number of mutations between the target pixel position and the related pixel position using the related pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position; binarizing the target image according to the number of mutations to obtain first information corresponding to the target image; determining a target area corresponding to the target pixel position, and calculating the entropy value information corresponding to the target pixel position in the target area to obtain second information; performing region segmentation on the target image according to the first information and the second information to obtain an initial area; performing similarity calculation on the initial area and a preset area, and then obtaining the target area corresponding to the initial area from the preset area.

[0059] For example, in order to better characterize the entropy information corresponding to the target pixel position, the local area can be adjusted from a square or rectangle to a circle, which can better match the natural shape, reduce edge effects, evenly cover the target area, simplify calculations, optimize feature extraction, and improve consistency and stability. These advantages help to more accurately characterize the entropy information corresponding to the target pixel position, thereby improving the effect of image analysis and anomaly detection.

[0060] Exemplarily, the target radius corresponding to the area circle in the image is determined based on expert experience, and the area circle corresponding to the target pixel position is determined based on the target radius and the target pixel position. Thus, by analyzing the pixels contained in the area circle, the associated related pixel positions are obtained, for example, each target pixel position is determined, and then based on the target radius, the pixel positions within the target radius are all considered to be related pixel positions.

[0061] Exemplarily, the pixel values ​​of the target pixel position and its associated related pixel positions are obtained. The number of mutations is determined by comparing the change between the target pixel value and its related pixel values. The number of mutations can be evaluated by calculating the amount of change in the pixel value (such as the change in grayscale value). If the change in the pixel value is large, the number of mutations is considered to be high.

[0062] Exemplarily, a threshold is selected, and when the number of mutations exceeds the threshold, the target pixel position is binarized, and the image is converted into a binary image, thereby obtaining the number of binarization processes corresponding to the target pixel position under the area circle, thereby obtaining the first information of the target image.

[0063] Exemplarily, entropy information of the target area is calculated, and then the entropy information is determined as the second information. Entropy information is an indicator for measuring the amount of information in the target area, and a high entropy value indicates that the information complexity in the target area is high. Entropy information can be calculated by analyzing the pixel distribution in the target area, which usually involves statistically analyzing the grayscale frequency distribution of pixels in the target area.

[0064] Exemplarily, the first information (from the binarization process) and the second information (the entropy value of the target area) are combined for region segmentation. This information can be used to refine the boundaries of the region segmentation to ensure the accuracy of the region segmentation. Based on the above information, an image segmentation algorithm (such as threshold segmentation, region growing, watershed algorithm, etc.) is applied to obtain the initial region.

[0065] Exemplarily, the preset region is a predefined region, usually a known standard region, which may be a known abnormal region, and then similarity is calculated between the initial region and the preset region using a similarity measurement method (such as intersection, overlap, shape similarity, etc.) to obtain a matching degree between the initial region and the preset region. Finally, based on the result of the similarity calculation, a target region that best matches the initial region is determined from the preset regions.

[0066] Specifically, by determining the associated pixel position of each target pixel position, the local area of ​​the target can be analyzed in detail, which helps to capture the local features and details of the target. The calculation of the number of mutations helps to identify significant differences between the target pixel and its surrounding pixels, which is very important for detecting boundaries, anomalies or feature changes. This provides support for subsequent improvement of the edge segmentation quality of the target image.

[0067] In some embodiments, the method of determining the number of mutations between the target pixel position and the related pixel position using the associated pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position comprises: after connecting the related pixel positions end to end, obtaining any two adjacent first positions and second positions between the related pixel positions, and obtaining first pixel information and second pixel information corresponding to the first positions from the associated pixel values; determining the number of mutations according to the first pixel information, the second pixel information and the target pixel value; wherein the number of mutations is obtained according to the following formula:

[0068] ;

[0069] ;

[0070] ;n represents any positive integer between 0, 1, 2, ..., sum-1;

[0071] in, Indicates the number of mutations corresponding to the target pixel position when the horizontal position is i and the vertical position is j, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, The function indicates that the output is 1 when the condition is met, and 0 when the condition is not met. Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, The target pixel value corresponding to the target pixel position when the horizontal position is i and the vertical position is j, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, represents the number of targets, and R represents the target radius.

[0072] Exemplarily, all associated related pixel positions are determined, and the related pixel positions are connected end to end according to the positions in the area circle to form a closed path, which ensures that all related pixel positions form a complete loop or a closed boundary.

[0073] Exemplarily, in a path connected end to end, any two adjacent pixel positions refer to continuous pixel points in the path. For example, if the pixel position sequence of the path is P1, P2, P3, ..., Pn, then adjacent pixel pairs include (P1, P2), (P2, P3), ..., (Pn-1, Pn) and (Pn, P1).

[0074] Exemplarily, for each pair of adjacent pixel positions (a first position and a second position), pixel information of these positions is extracted from the associated pixel values. For example, first pixel information of the first position and second pixel information of the second position are obtained.

[0075] Exemplarily, the number of mutations refers to the number of times the pixel values ​​between adjacent pixel positions change significantly. The number of mutations is obtained using the first pixel information, the second pixel information and the target pixel value according to the following formula:

[0076] ;

[0077] ;

[0078] ;n represents any positive integer between 0, 1, 2, ..., sum-1;

[0079] in, Indicates the number of mutations corresponding to the target pixel position when the horizontal position is i and the vertical position is j. Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position when The function indicates that the output is 1 when the condition is met, and 0 when the condition is not met. Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position when The target pixel value corresponding to the target pixel position when the horizontal position is i and the vertical position is j, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position when Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position when represents the number of targets, and R represents the target radius.

[0080] For example, using the above formula and Substitute the corresponding n value to obtain the horizontal and vertical coordinates at the corresponding n value, where n represents any positive integer between 0, 1, 2, ..., sum-1. For example, substitute n equals sum-1 into and Then the horizontal position can be obtained as , the vertical position is .

[0081] For example, the number of mutations corresponding to the target pixel position under the related pixel position is calculated according to the above formula, so as to evaluate the change characteristics between the target pixel position and its related pixel position, and then effectively feedback whether the target pixel position is the boundary point of the target area, or reflects the texture complexity and uniformity of the target area. This feedback is of great significance for image analysis and processing, and can help further boundary detection.

[0082] In some embodiments, the target image is segmented into regions according to the first information and the second information to obtain an initial region, including: performing a ratio operation on the first information and the second information to obtain edge representation information corresponding to the target pixel position; comparing the edge representation information with preset information to determine an edge curve corresponding to the target image; and the target image is segmented into regions according to the edge curve to obtain the initial region.

[0083] Exemplarily, a ratio operation is performed on the first information (e.g., the result of the binary image) and the second information (e.g., the entropy value of the target area). Specifically, for each target pixel position, a ratio is calculated based on the first information and the second information corresponding to it, and then the ratio is determined as edge representation information, which indicates that the target pixel position corresponds to a larger change or has a more significant edge feature.

[0084] Exemplarily, the preset information may be a reference edge feature or a known edge template. This information is used to help identify and verify the edge in the target image, and then compare the obtained edge representation information with the preset information. This can be done by calculating the size between the edge representation information and the preset information. When the edge representation information is greater than the preset information, the target pixel position is determined as a pixel point in the edge curve, and then all the pixel points are obtained, so that the pixel points are fitted to obtain the edge curve. The edge curve refers to the contour line or edge path of the abnormal area in the target image.

[0085] Exemplarily, a determined edge curve is used to guide the image region segmentation. The edge curve provides clear boundary information for the segmentation. Thus, the target image is divided into different regions according to the edge curve. Usually, these regions are divided with the edge curve as the boundary to form an initial segmentation result, that is, multiple initial regions are obtained.

[0086] Specifically, by calculating the ratio of the first information to the second information, the effect of edge detection can be enhanced. This comparison can usually improve the accuracy of detection, especially when the target pixel position has specific edge features. Then, by comparing the obtained edge representation information with the preset information, the correctness of the edge curve can be verified at different scales and directions to ensure the consistency and repeatability of edge detection. Finally, using the determined edge curve for region segmentation can improve the accuracy of the segmentation result. The edge curve can accurately track the boundary of the target object or region, which helps to achieve a more accurate segmentation effect.

[0087] In some embodiments, performing similarity calculation on the initial area and the preset area to obtain the target area corresponding to the initial area from the preset area includes: performing standardized cutting on the initial area to obtain a standard area corresponding to the initial area, wherein the standard area is equal in size to the preset area; calculating the pixel difference between the standard area and the preset area to obtain a similarity result; and obtaining the target area corresponding to the initial area from the preset area based on the similarity result.

[0088] Exemplarily, the preset region is an area with a specific size and shape, which will be used as a reference for comparison. Make sure that the exact size of the preset region is known, and then use image processing techniques (such as cropping, scaling, etc.) to adjust the initial region to the same size as the preset region. This can be achieved by interpolation algorithms (such as nearest neighbor interpolation, bilinear interpolation, etc.) to achieve size matching, that is, the cut area is the standard area, which will be used for subsequent comparison with the preset region.

[0089] For example, ensure that the standard area and the preset area are aligned before comparison, that is, their corresponding pixels represent the same position in the image. The purpose of segmenting the initial area into standard areas is to solve the problem of the same material but inconsistent shape in the plate, so that the material of the initial area can be accurately identified when the segmented standard area is compared with the preset area, thereby providing support for subsequent anomaly detection.

[0090] Exemplarily, the pixel values ​​of the standard area and the preset area are compared pixel by pixel. For each pixel position, the difference (which may be an absolute value difference or other difference metric) of the corresponding pixel values ​​in the two areas is calculated, thereby generating a difference map of the same size as the standard area and the preset area, in which each pixel value represents the degree of difference at the corresponding position. The difference map is then evaluated to determine the overall similarity between the standard area and the preset area. Various methods (such as average difference, difference histogram, etc.) can be used to quantify the similarity.

[0091] Exemplarily, based on the similarity result, the part that best matches the standard area is determined from the preset area as the target area. This can be achieved by selecting the area with the smallest difference in the difference map. Finally, based on the similarity result, the target area corresponding to the initial area is extracted from the preset area. That is, the plate characteristics corresponding to the initial area are obtained.

[0092] Step S104: Perform outlier detection on the initial sensor data to obtain target sensor data.

[0093] For example, the initial sensor data may include various sensor readings, such as temperature, humidity, pressure, etc., and the initial sensor data is processed for noise and inconsistency to ensure the quality of the data. Thus, statistical methods, Z-score methods or outlier detection based on box plots, or machine learning methods or deep learning methods are used to establish an outlier detection model to perform outlier detection on the initial sensor data, thereby obtaining outliers in the initial sensor data and determining them as target sensor data.

[0094] In addition, before performing outlier detection on the initial sensor data, it is also necessary to classify the collected initial sensor data to obtain real-time production data and generate the corresponding order panels, that is, the processing data corresponding to the target object, process completion status, equipment capacity, video monitoring data and other data. The above data are integrated and analyzed to form unified production real-time data (panel processing quality data, capacity data, equipment status data, real-time data of key workstations), so as to identify outliers based on the unified production real-time data.

[0095] In some embodiments, performing outlier detection on the initial sensor data to obtain target sensor data includes: obtaining adjacent sensor data corresponding to any one selected sensor data in the initial sensor data; determining the same number of identical sensor data corresponding to the adjacent sensor data corresponding to any two of the selected sensor data; determining the data similarity between the two selected sensor data based on the same number; calculating the discrete probability corresponding to the selected sensor data based on the data similarity, and performing outlier detection based on the discrete probability to obtain the target sensor data.

[0096] Exemplarily, one selected sensor data (referred to as selected sensor data A) is randomly determined from the initial sensor data, and then adjacent sensor data with a close distance to the selected sensor data A is searched. Assuming that the selected sensor data A and the selected sensor data B are selected, the adjacent sensor data corresponding to the selected sensor data A and the selected sensor data B are respectively extracted.

[0097] Exemplarily, the adjacent sensor data of the selected sensor data A and the selected sensor data B are compared to find the number of identical sensor data between them. For example, if the adjacent sensor data of the selected sensor data A are A1 and A2, and the adjacent sensor data of the selected sensor data B are A1, B1, and B2, then the number of identical sensor data corresponding to the existence of identical sensor data between the two adjacent sensor data is calculated.

[0098] Exemplarily, the data similarity between the selected sensor data A and the selected sensor data B is calculated based on the number of identical sensor data found in the previous step. First, the number of identical data between the selected sensor data A and any one of the remaining initial sensor data is determined, and then the identical data is screened according to a preset number. When the identical data is greater than or equal to the preset number, the data similarity is calculated based on the screened identical data. For example, the distance between the selected sensor data A and the associated sensor data corresponding to the screened identical data is calculated, and then the weighted similarity between the two is determined based on the identical data divided by the distance, and then the weighted similarities corresponding to all the screened associated sensor data corresponding to the identical data are summed and averaged to obtain the data similarity.

[0099] Exemplarily, the first data similarity corresponding to each selected sensor data is finally obtained, and the second data similarity corresponding to the associated sensor data when the same quantity is filtered is obtained, and then all the second data similarities are summed to obtain the total similarity value, and then the quantity corresponding to the second data similarities is obtained, and then the first data similarity is multiplied by the quantity and divided by the total similarity value to obtain the discrete probability corresponding to the selected sensor data, which reflects the distribution of the selected sensor data.

[0100] Exemplarily, outlier detection is performed based on discrete probability. A threshold can usually be set. If the discrete probability is higher than the threshold, it is considered that the selected sensor data is abnormal, and the selected sensor data corresponding to the discrete probability is determined as the target sensor data. The purpose of outlier detection is to identify sensor data points that are significantly different from other data points.

[0101] Specifically, analyzing the data similarity between the selected sensor data and its adjacent sensor data helps to discover abnormal patterns in local data, and then calculating discrete probabilities based on similarity can better identify outliers that are inconsistent with other sensor data. Determining data similarity can reveal the local consistency of sensor data and identify potential data anomalies.

[0102] Step S105: fusing the target area and the target sensor data to obtain a data governance result corresponding to the target object.

[0103] Exemplarily, the target region represents an abnormal region in the image detection stage, and the target sensor data represents abnormal information in the sensor detection stage.

[0104] Exemplarily, the mapping sensor data corresponding to the target area is obtained from the initial sensor data, and the mapping sensor data and the target sensor data are intersected, and the data management result corresponding to the target object is: when there are more intersection results, it indicates that the target area is an abnormal area; when there are fewer intersection results, it indicates that the target area is not an abnormal area, and it is necessary to determine the area corresponding to the target sensor data in the target image as an abnormal area.

[0105] In some embodiments, the information fusion of the target area and the target sensor data to obtain the data governance result corresponding to the target object includes: obtaining the associated area corresponding to the target sensor data in the target image according to the monitoring position corresponding to the target sensor data; comparing the difference between the associated area and the target area, and determining the data governance result corresponding to the target object according to the difference.

[0106] For example, during the image detection stage, it can be found that there are color anomalies, scratches, etc. on the surface of the plate. During the sensor detection stage, it can be found that the material of the plate is abnormal. Combining the two can reveal problems at multiple angles of the target object.

[0107] Exemplarily, the monitoring position of the sensor monitoring the target object corresponding to the target sensor data is obtained, and then the monitoring position is converted to the corresponding associated area in the target image, that is, the target sensor data is associated with the position information in the target image.

[0108] Exemplarily, the differences between the associated area and the target area are compared. When the associated area is consistent with the target area, the data management result is that the plate information corresponding to the target area has surface abnormalities and sensor abnormalities, that is, the plate corresponding to the target area has both quality and color problems; when the associated area is inconsistent with the target area, the data management result is that the plate information corresponding to the target area has surface abnormalities, but the material is normal, and the plate information corresponding to the associated area has normal surface but abnormal material.

[0109] Specifically, by distinguishing between surface anomalies and material anomalies, specific defects in the sheet can be more accurately identified, so that targeted repair or replacement measures can be taken, which can reduce rework or resource waste caused by misjudgment and improve the efficiency and accuracy of fault detection. By accurately detecting and solving the quality and color problems of the sheet, the quality of the final product can be significantly improved to meet customer expectations. Therefore, through detailed difference analysis, data governance can be more accurately carried out to ensure that all anomalies are identified and processed. This can significantly improve the quality control and anomaly detection capabilities of the production process, thereby achieving more efficient and accurate production and maintenance management.

[0110] In addition, after the target area and target sensor data are fused to obtain the data governance results corresponding to the target object, it is necessary to perform feature analysis on the collected target area and target sensor data to distinguish the abnormal types corresponding to the target object, such as abnormal plate processing quality, equipment abnormalities, and abnormalities in the key workstation video area, so as to distribute the different abnormal types detected to different personnel for processing. Therefore, according to the abnormal type detected, the on-site personnel are notified by email or sound and light to confirm the abnormality in time to quickly handle various abnormalities.

[0111] The embodiment of the present invention provides a data management method and system for an intelligent plate processing factory, the method comprising acquiring images at different points of interest and obtaining sensor data at the same time, so as to obtain information of a target object from multiple angles. This all-round data collection ensures comprehensive monitoring of the target object and reduces the problem that may be missed by a single perspective. Then, the first image and the second image are subjected to information fusion to generate a target image. This fusion can integrate the details of different points of interest, provide a more comprehensive view of the target object, and help to more accurately identify and analyze potential problem areas. By applying information entropy and binarization processing to the target image, anomaly detection is performed. This process helps to accurately identify abnormal areas in the target image. Information entropy analysis provides a measure of the amount of information in the image, while binarization processing simplifies the identification of abnormal areas and effectively improves the accuracy of anomaly detection. Then, anomaly detection is performed on the initial sensor data, and the abnormality of the target object under sensor acquisition can be found, thereby finding the abnormality of the target object under sensor acquisition, and finally the target area is fused with the target sensor data to obtain the data management result of the target object. This comprehensive analysis can more accurately evaluate the state of the target object, thereby providing more comprehensive quality control and anomaly monitoring. This can effectively improve the monitoring and anomaly detection capabilities of target objects in the sheet metal processing process, ensure timely discovery and accurate analysis of problems, and thus improve product quality, reduce production costs, and optimize production efficiency. It also solves the problem that related technologies often manage and analyze sheet metal processing data from only a single perspective, which greatly increases the risk of misjudgment and causes potential problems to be unable to be discovered in a timely and accurate manner.

[0112] See also Figure 2 , Figure 2 A data management system 200 for an intelligent sheet metal processing factory is provided in an embodiment of the present application. The data management system 200 for the intelligent sheet metal processing factory includes a data acquisition module 201, an information fusion module 202, a region segmentation module 203, an anomaly monitoring module 204, and a result determination module 205, wherein the data acquisition module 201 is used to acquire a first image corresponding to a target object at a first focus point and a second image corresponding to a second focus point during sheet metal processing, and simultaneously obtain initial sensor data corresponding to the target object under a target sensor; the information fusion module 202 is used to perform information fusion on the first image and the second image to obtain a target image; the region segmentation module 203 is used to perform anomaly detection on the target image using information entropy and binarization processing to obtain a target region corresponding to the target image; the anomaly monitoring module 204 is used to perform anomaly value detection on the initial sensor data to obtain target sensor data; the result determination module 205 is used to perform information fusion on the target region and the target sensor data to obtain a data management result corresponding to the target object.

[0113] In some implementations, the information fusion module 202 performs, during the process of fusing the first image and the second image to obtain the target image:

[0114] Extracting texture features from the first image to obtain a first texture image and extracting texture features from the second image to obtain a second texture image;

[0115] Determine a first target window, obtain a first partial image corresponding to the first image in the first target window, and obtain a second partial image corresponding to the first texture image in the first target window;

[0116] Determine a first pixel mean corresponding to a first pixel value in the first partial image and determine a second pixel mean and a first pixel variance corresponding to a second pixel value in the second partial image;

[0117] Determine a first blurriness, and establish a first pixel association relationship corresponding to the first image and the first texture image in the first target window according to the first blurriness, the first pixel mean, the second pixel mean, and the first pixel variance;

[0118] Perform pixel enhancement on the first image according to the first pixel association relationship to obtain a first enhanced image;

[0119] Determine a second target window, obtain a third partial image corresponding to the second image in the second target window, and obtain a fourth partial image corresponding to the second texture image in the second target window;

[0120] Determine a third pixel mean corresponding to a third pixel value in the third partial image and determine a fourth pixel mean and a second pixel variance corresponding to a fourth pixel value in the fourth partial image;

[0121] Determine a second blurriness, and establish a second pixel association relationship corresponding to the second image and the second texture image in the second target window according to the second blurriness, the third pixel mean, the fourth pixel mean, and the second pixel variance;

[0122] Performing pixel enhancement on the second image according to the second pixel association relationship to obtain a second enhanced image;

[0123] The target image is obtained by performing information fusion based on the first enhanced image and the second enhanced image.

[0124] In some implementations, the information fusion module 202, in the process of fusing the information according to the first enhanced image and the second enhanced image to obtain the target image, executes:

[0125] Comparing the first enhanced image with the first image pixel by pixel according to a first bitwise comparison algorithm to obtain a first comparison image;

[0126] Comparing the second enhanced image and the second image pixel by pixel according to a second bitwise comparison algorithm to obtain a second comparison image, wherein the comparison methods of the first bitwise comparison algorithm and the second bitwise comparison algorithm are opposite;

[0127] The first comparison image and the second comparison image are merged and denoised to obtain a target comparison image;

[0128] The first image and the second image are fused according to the target contrast image to obtain the target image.

[0129] In some implementations, the region segmentation module 203 performs, during the process of performing abnormality detection on the target image by using information entropy and binarization processing to obtain the target region corresponding to the target image:

[0130] Determine the target radius corresponding to the area circle and the number of targets corresponding to the area circle;

[0131] Obtaining a relevant pixel position associated with each target pixel position in the target image according to the target radius and the target quantity;

[0132] Determine the number of mutations between the target pixel position and the related pixel position using the associated pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position;

[0133] Performing binarization processing on the target image according to the number of mutations to obtain first information corresponding to the target image;

[0134] Determine a target area corresponding to the target pixel position, and calculate entropy value information corresponding to the target pixel position in the target area to obtain second information;

[0135] Performing region segmentation on the target image according to the first information and the second information to obtain an initial region;

[0136] The initial area and the preset area are similarly calculated, and then the target area corresponding to the initial area is obtained from the preset area.

[0137] In some implementations, the region segmentation module 203, in the process of determining the number of mutations between the target pixel position and the related pixel position using the associated pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position, performs:

[0138] After connecting the related pixel positions end to end, obtaining any two adjacent first positions and second positions between the related pixel positions, and obtaining first pixel information and second pixel information corresponding to the first positions from the associated pixel values;

[0139] Determine the number of mutations according to the first pixel information, the second pixel information and the target pixel value;

[0140] Wherein, the number of mutations is obtained according to the following formula:

[0141] ;

[0142] ;

[0143] ;n represents any positive integer between 0, 1, 2, ..., sum-1;

[0144] in, Indicates the number of mutations corresponding to the target pixel position when the horizontal position is i and the vertical position is j, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, The function indicates that the output is 1 when the condition is met, and 0 when the condition is not met. Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, The target pixel value corresponding to the target pixel position when the horizontal position is i and the vertical position is j, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, Indicates the horizontal position is , the vertical position is The pixel value corresponding to the relevant pixel position corresponding to the time, represents the number of targets, and R represents the target radius.

[0145] In some implementations, the region segmentation module 203, in the process of segmenting the target image according to the first information and the second information to obtain the initial region, executes:

[0146] Performing a ratio operation on the first information and the second information to obtain edge representation information corresponding to the target pixel position;

[0147] Comparing the edge representation information with preset information to determine an edge curve corresponding to the target image;

[0148] The target image is segmented into regions according to the edge curve to obtain the initial region.

[0149] In some implementations, the region segmentation module 203 performs, in the process of performing similarity calculation on the initial region and the preset region to obtain the target region corresponding to the initial region from the preset region:

[0150] Performing standardized cutting on the initial area to obtain a standard area corresponding to the initial area, wherein the standard area is equal in size to the preset area;

[0151] Calculating the pixel difference between the standard area and the preset area to obtain a similarity result;

[0152] The target area corresponding to the initial area is obtained from the preset area according to the similarity result.

[0153] In some implementations, the abnormality monitoring module 204 performs abnormal value detection on the initial sensor data to obtain target sensor data, including:

[0154] Obtaining adjacent sensor data corresponding to any selected sensor data in the initial sensor data;

[0155] Determine the same number of identical sensor data corresponding to any two adjacent sensor data corresponding to the selected sensor data;

[0156] Determine the data similarity between the two selected sensor data according to the same number;

[0157] The discrete probability corresponding to the selected sensor data is calculated according to the data similarity, and outlier detection is performed according to the discrete probability to obtain the target sensor data.

[0158] In some implementations, the result determination module 205 performs, during the process of fusing the target area and the target sensor data to obtain the data governance result corresponding to the target object:

[0159] Obtaining, according to the monitoring position corresponding to the target sensor data, an associated area corresponding to the target image of the target sensor data;

[0160] The difference between the associated area and the target area is compared, and the data governance result corresponding to the target object is determined according to the difference.

[0161] In some embodiments, the data governance system 200 of the smart sheet metal processing factory can be applied to a terminal device.

[0162] It should be noted that technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of the data governance system 200 of the intelligent plate processing factory described above can refer to the corresponding process in the aforementioned data governance method embodiment of the intelligent plate processing factory, and will not be repeated here.

[0163] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any data governance method for an intelligent plate processing factory provided in the description of the embodiment of the present invention.

[0164] The storage medium may be an internal storage unit of the terminal device in the aforementioned embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device.

[0165] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0166] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0167] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A data management method for an intelligent sheet metal processing factory, characterized in that: The method comprises: Collecting a first image corresponding to a first focus point and a second image corresponding to a second focus point of a target object during plate processing, and simultaneously obtaining initial sensor data corresponding to the target object under a target sensor; Performing information fusion on the first image and the second image to obtain a target image; Performing abnormality detection on the target image by using information entropy and binarization processing to obtain a target area corresponding to the target image; Performing outlier detection on the initial sensor data to obtain target sensor data; Performing information fusion on the target area and the target sensor data to obtain a data governance result corresponding to the target object; The step of fusing the first image and the second image to obtain a target image includes: Extracting texture features from the first image to obtain a first texture image and extracting texture features from the second image to obtain a second texture image; Determine a first target window, obtain a first partial image corresponding to the first image in the first target window, and obtain a second partial image corresponding to the first texture image in the first target window; Determine a first pixel mean corresponding to a first pixel value in the first partial image and determine a second pixel mean and a first pixel variance corresponding to a second pixel value in the second partial image; Determine a first blurriness, and establish a first pixel association relationship corresponding to the first image and the first texture image in the first target window according to the first blurriness, the first pixel mean, the second pixel mean, and the first pixel variance; Perform pixel enhancement on the first image according to the first pixel association relationship to obtain a first enhanced image; Determine a second target window, obtain a third partial image corresponding to the second image in the second target window, and obtain a fourth partial image corresponding to the second texture image in the second target window; Determine a third pixel mean corresponding to a third pixel value in the third partial image and determine a fourth pixel mean and a second pixel variance corresponding to a fourth pixel value in the fourth partial image; Determine a second blurriness, and establish a second pixel association relationship corresponding to the second image and the second texture image in the second target window according to the second blurriness, the third pixel mean, the fourth pixel mean, and the second pixel variance; Performing pixel enhancement on the second image according to the second pixel association relationship to obtain a second enhanced image; The target image is obtained by performing information fusion based on the first enhanced image and the second enhanced image.

2. The method according to claim 1, characterized in that The step of fusing information according to the first enhanced image and the second enhanced image to obtain the target image includes: Comparing the first enhanced image with the first image pixel by pixel according to a first bitwise comparison algorithm to obtain a first comparison image; Comparing the second enhanced image and the second image pixel by pixel according to a second bitwise comparison algorithm to obtain a second comparison image, wherein the comparison methods of the first bitwise comparison algorithm and the second bitwise comparison algorithm are opposite; The first comparison image and the second comparison image are merged and denoised to obtain a target comparison image; The first image and the second image are fused according to the target contrast image to obtain the target image.

3. The method according to claim 1, characterized in that The method of performing abnormality detection on the target image by using information entropy and binarization processing to obtain a target area corresponding to the target image includes: Determine the target radius corresponding to the area circle and the number of targets corresponding to the area circle; Obtaining a relevant pixel position associated with each target pixel position in the target image according to the target radius and the target quantity; Determine the number of mutations between the target pixel position and the related pixel position using the associated pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position; Performing binarization processing on the target image according to the number of mutations to obtain first information corresponding to the target image; Determine a target area corresponding to the target pixel position, and calculate entropy value information corresponding to the target pixel position in the target area to obtain second information; Performing region segmentation on the target image according to the first information and the second information to obtain an initial region; The initial area and the preset area are similarly calculated, and then the target area corresponding to the initial area is obtained from the preset area.

4. The method according to claim 3, characterized in that The determining the number of mutations between the target pixel position and the related pixel position by using the associated pixel value corresponding to the related pixel position and the target pixel value corresponding to the target pixel position comprises: After connecting the related pixel positions end to end, obtaining any two adjacent first positions and second positions between the related pixel positions, and obtaining first pixel information and second pixel information corresponding to the first positions from the associated pixel values; Determine the number of mutations according to the first pixel information, the second pixel information and the target pixel value; Wherein, the number of mutations is obtained according to the following formula: x n =-R*sin(2*pi*n / sum); y n =R*cos(2*pi*n / sum); n represents any positive integer between 0, 1, 2, ..., sum-1; Among them, T ij represents the number of mutations corresponding to the target pixel position when the horizontal position is i and the vertical position is j, P(x sum-1 ,y sum-1 ) indicates the horizontal position is x sum-1 , the vertical position is y sum-1 The IF function indicates that the output is 1 when the condition is met and 0 when the condition is not met. P(x0, y0) indicates the pixel value corresponding to the relevant pixel position when the horizontal position is x0 and the vertical position is y0. P(i, j) indicates the target pixel value corresponding to the target pixel position when the horizontal position is i and the vertical position is j. P(x0, y0) indicates the pixel value corresponding to the relevant pixel position when the horizontal position is x0 and the vertical position is y0. t ,y t ) indicates the horizontal position is x t , the vertical position is y t The pixel value corresponding to the relevant pixel position at the time, P(x t-1 ,y t-1 ) indicates the horizontal position is x t-1 , the vertical position is y t-1 The pixel value corresponding to the relevant pixel position at the time, sum represents the number of targets, R represents the target radius, IF(P(x sum-1 ,y sum-1 )>P(i, j), 1, 0) indicates that the horizontal position is x sum-1 , the vertical position is y sum-1 The pixel value P(x sum-1 ,y sum-1 ) is greater than the target pixel value P(i, j) corresponding to the target pixel position corresponding to the horizontal position i and the vertical position j, IF(P(x sum-1 ,y sum-1 )>P(i,j),1,0) is equal to 1, otherwise IF(P(x sum-1 ,y sum-1 )>P(i,j),1,0) is equal to 0; IF(P(x0,y0)>P(i,j),1,0) means that the pixel value P(x0,y0) corresponding to the relevant pixel position when the horizontal position is x0 and the vertical position is y0 is greater than the target pixel value P(i,j) corresponding to the target pixel position when the horizontal position is i and the vertical position is j, IF(P(x0,y0)>P(i,j),1,0) is equal to 1, otherwise IF(P(x0,y0)>P(i,j),1,0) is equal to 0; IF(P(x0,y0)>P(i,j),1,0) means that the pixel value P(x0,y0) corresponding to the relevant pixel position when the horizontal position is x0 and the vertical position is y0 is greater than the target pixel value P(i,j) corresponding to the target pixel position when the horizontal position is i and the vertical position is j, t ,y t )>P(i, j), 1, 0) indicates that the horizontal position is x t , the vertical position is y t The pixel value P(x t ,y t ) is greater than the target pixel value P(i, j) corresponding to the target pixel position corresponding to the horizontal position i and the vertical position j. t ,y t )>P(i,j),1,0) is equal to 1, otherwise IF(P(x t ,y t )>P(i, j), 1, 0) is equal to 0; IF(P(x t-1 ,y t-1 )>P(i, j), 1, 0) indicates that the horizontal position is x t-1 , the vertical position is y t-1 The pixel value P(x t-1 ,y t-1 ) is greater than the target pixel value P(i, j) corresponding to the target pixel position corresponding to the horizontal position i and the vertical position j. t-1 ,y t-1 )>P(i,j),1,0) is equal to 1, otherwise IF(P(x t-1 ,y t-1 )>P(i,j),1,0) is equal to 0,|IF(P(x sum-1 ,y sum-1 )>P(i,j),1,0)-IF(P(x0,y0)>P(i,j),1,0)|means IF(P(x0,y0)>P(i,j),1,0) sum-1 ,y sum-1 )>P(i,j),1,0) and IF(P(x0,y0)>P(i,j),1,0),|IF(P(x0,y0)>P(i,j),1,0) t ,y t )>P(i,j),1,0)-IF(P(x t-1 ,y t-1 )>P(i, j), 1, 0)| means IF(P(x t ,y t )>P(i,j),1,0) and IF(P(x t-1 ,y t-1 )>P(i, j), 1, 0).

5. The method according to claim 3, characterized in that: The performing region segmentation on the target image according to the first information and the second information to obtain an initial region includes: Performing a ratio operation on the first information and the second information to obtain edge representation information corresponding to the target pixel position; Comparing the edge representation information with preset information to determine an edge curve corresponding to the target image; The target image is segmented into regions according to the edge curve to obtain the initial region.

6. The method according to claim 3, characterized in that The performing similarity calculation on the initial area and the preset area, and then obtaining the target area corresponding to the initial area from the preset area, includes: Performing standardized cutting on the initial area to obtain a standard area corresponding to the initial area, wherein the standard area is equal in size to the preset area; Calculating the pixel difference between the standard area and the preset area to obtain a similarity result; The target area corresponding to the initial area is obtained from the preset area according to the similarity result.

7. The method according to claim 1, characterized in that The performing outlier detection on the initial sensor data to obtain target sensor data includes: Obtaining adjacent sensor data corresponding to any selected sensor data in the initial sensor data; Determine the same number of identical sensor data corresponding to any two adjacent sensor data corresponding to the selected sensor data; Determine the data similarity between the two selected sensor data according to the same number; The discrete probability corresponding to the selected sensor data is calculated according to the data similarity, and outlier detection is performed according to the discrete probability to obtain the target sensor data.

8. The method according to claim 1, characterized in that The step of fusing the target area and the target sensor data to obtain a data governance result corresponding to the target object includes: Obtaining, according to the monitoring position corresponding to the target sensor data, an associated area corresponding to the target image of the target sensor data; The difference between the associated area and the target area is compared, and the data governance result corresponding to the target object is determined according to the difference.

9. A data management system for an intelligent sheet metal processing factory, characterized in that: include: A data acquisition module, used to acquire a first image corresponding to a first focus point and a second image corresponding to a second focus point of a target object during plate processing, and simultaneously obtain initial sensor data corresponding to the target object under a target sensor; An information fusion module is used to perform information fusion on the first image and the second image to obtain a target image; wherein, the information fusion on the first image and the second image to obtain the target image includes: extracting texture features on the first image to obtain a first texture image and extracting texture features on the second image to obtain a second texture image; determining a first target window to obtain a first local image corresponding to the first image in the first target window and a second local image corresponding to the first texture image in the first target window; determining a first pixel mean corresponding to a first pixel value in the first local image and determining a second pixel mean and a first pixel variance corresponding to a second pixel value in the second local image; determining a first blur, and establishing a first blur corresponding to the first image and the first texture image under the first target window according to the first blur, the first pixel mean, the second pixel mean and the first pixel variance. a pixel association relationship; performing pixel enhancement on the first image according to the first pixel association relationship to obtain a first enhanced image; determining a second target window, obtaining a third local image corresponding to the second image in the second target window and obtaining a fourth local image corresponding to the second texture image in the second target window; determining a third pixel mean corresponding to a third pixel value in the third local image and determining a fourth pixel mean and a second pixel variance corresponding to a fourth pixel value in the fourth local image; determining a second blur, and establishing a second pixel association relationship between the second image and the second texture image under the second target window according to the second blur, the third pixel mean, the fourth pixel mean and the second pixel variance; performing pixel enhancement on the second image according to the second pixel association relationship to obtain a second enhanced image; performing information fusion based on the first enhanced image and the second enhanced image to obtain the target image; A region segmentation module is used to perform abnormality detection on the target image by using information entropy and binarization processing to obtain a target region corresponding to the target image; An abnormality monitoring module, used for performing abnormal value detection on the initial sensor data to obtain target sensor data; A result determination module is used to fuse the target area and the target sensor data to obtain the data management result corresponding to the target object.