Industrial internet quality detection system based on edge cloud collaboration

Abstract

The invention relates to the field of industrial internet, in particular to an industrial internet quality detection system based on edge cloud collaboration, which is characterized in that an edge acquisition module acquires image data of a production line according to a preset time period, and an edge image processing module extracts contrast features and surface texture features in the image data; a defect area contour is identified based on a standard product contour, an edge image analysis module calculates a surface anomaly index and a defect anomaly index, and a comprehensive defect score is obtained through weighting; the edge uploading module sets labels for the image data according to the comprehensive defect scores, sampling sending under high-quality labels and complete sending under low-quality labels are achieved, and the cloud analysis module conducts high-precision defect evaluation on the received image data and conducts fault early warning through time sequence analysis. The method improves the detection efficiency, reduces the network bandwidth pressure through a hierarchical uploading mechanism, and achieves the judgment of the quality degradation trend through the high-precision evaluation and time sequence analysis capabilities of the cloud.

Description

Technical Field

[0001] This invention relates to the field of industrial internet, and in particular to an industrial internet quality inspection system based on edge cloud collaboration. Background Technology

[0002] As a crucial pillar of the national economy, industrial manufacturing directly impacts a company's core competitiveness and market reputation through product quality. During continuous production, industrial parts and product surfaces often exhibit defects such as scratches, dents, rust, and stains due to processing deviations, equipment wear, or environmental factors. Furthermore, the rapid production cycle and large volume of inspections make traditional manual visual inspection methods insufficient for meeting the demands of high-efficiency, high-precision quality control. To improve product quality consistency and ensure stable production line operation, automated and intelligent quality inspection systems are needed. Currently, machine vision and industrial internet technologies are commonly used to digitize and intelligentize the inspection process.

[0003] For example, Chinese Patent Publication No. CN114677239B discloses a critical component error-proofing detection system based on edge computing, relating to the field of industrial internet technology. The workflow of this invention is as follows: After installation, the system reaches the monitoring point; the system automatically reads the barcode SN and the critical component QR code; the system requests the MES system to retrieve the order and BOM information corresponding to the SN; the MES system transmits the order and BOM, and the critical component's theoretical material code DBOM; if the system compares the QR and DBOM and they are equal, the critical component's qualified result is pushed to the MES, allowing it to proceed to the next process; if they are not equal, the critical component's unqualified result is pushed to the MES, and it is not allowed to proceed to the next process. This invention, through industrial internet technology, connects multiple systems for cloud-edge collaboration in edge computing, enabling automatic synchronization of the correspondence between newly added material codes for critical components and the critical component collection points. It also utilizes industrial cameras with innovative hardware filtering and software code filtering technologies to achieve a closed-loop process quality control in the cloud, improving factory error-proofing efficiency. However, the existing technology still has the following problems. 1. In existing technologies, cloud computing capabilities are heavily relied upon. In a large number of distributed scenarios involving multiple product lines, the reliance on cloud scheduling for a large number of edge detection nodes results in low detection efficiency, making it difficult to meet the real-time quality judgment requirements of high-speed production lines. 2. In the existing technology, no hierarchical data transmission mechanism has been established to upload all collected data completely to the cloud, resulting in excessively large data transmission flow and significant network bandwidth pressure. 3. In existing technologies, quality inspection often focuses on identifying single-point defects, and lacks the ability to determine the process deviations and quality deterioration trends that gradually accumulate during the production process. Summary of the Invention

[0004] To address these issues, this invention provides an industrial internet quality inspection system based on edge cloud collaboration. This system overcomes the problems of existing technologies that rely heavily on cloud computing capabilities, lack a hierarchical data transmission mechanism, and often focus on single-point defect identification, resulting in low inspection efficiency, high network bandwidth pressure, and a lack of judgment on the gradually accumulating process deviations and quality deterioration trends during production.

[0005] To achieve the above objectives, the present invention provides an industrial internet quality inspection system based on edge cloud collaboration, comprising: The edge acquisition module is used to acquire image data of the production line according to a preset time period; An edge image processing module, which is connected to the edge acquisition module, is used to extract contrast features and surface texture features from image data and identify the contour of defect areas based on a preset standard product contour. An edge image analysis module, which is connected to the edge image processing module, is used to calculate a surface anomaly index based on contrast features and surface texture features, calculate a defect anomaly index based on the defect region contour, and determine the weighted result of the surface anomaly index and the defect anomaly index as a comprehensive defect score. An edge upload module, connected to the edge image analysis, is used to upload image data based on a comprehensive defect score, including: Labels are assigned to the image data corresponding to the time period based on the comprehensive defect score; In response to a high-quality tag, the image data is sampled and sent at preset intervals; Alternatively, in response to a low-quality tag, send the complete image data; The cloud analysis module, which is connected to the edge upload module, is used to perform high-precision defect assessment based on the received image data, and to perform time-series analysis on the high-precision defect assessment results in order to provide early warning of faults based on the changing trend of the high-precision defect assessment results.

[0006] Furthermore, the edge image processing module is used to identify the contour of the defect region based on a preset standard product contour, including, Identify the product outline of a target product in image data; The product outline is superimposed and compared with a preset standard product outline to identify the difference areas where the product outline and the standard outline differ. All discrepancy regions in the edge image are identified as the defect region contours.

[0007] Furthermore, the edge image processing module overlays and compares the product outline with a preset standard product outline to identify the difference regions where the product outline and the standard outline differ, including... The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several blocks. Identify the areas where there are differences between the product outline and the standard product outline in the overlay comparison image, and filter out the difference blocks that meet the preset difference criteria; The difference blocks are connected and merged to form several closed difference regions.

[0008] Furthermore, the edge image analysis module is used to calculate the surface anomaly index based on contrast features and surface texture features, including: The ratio of the contrast characteristic value to the contrast standard value is defined as the contrast index; The ratio of surface texture feature values ​​to texture standard values ​​is defined as the texture index; The surface anomaly index is obtained by weighted summation of the contrast index and the texture index.

[0009] Furthermore, the edge image analysis module is used to calculate a defect anomaly index based on the defect region contour, including: The ratio of the total area of ​​the defective region outline to the area of ​​the standard product is defined as the area index; The ratio of the number of differential regions in the defect region outline to the number of standard defects is determined as the quantity index; The defect anomaly index is obtained by weighted summation of the area index and the quantity index.

[0010] Furthermore, the edge upload module sets labels for the image data corresponding to the time period based on the comprehensive defect score, including: If the overall defect score is less than the quality threshold, then a high-quality label is set for the image data corresponding to the time period. If the overall defect score is greater than or equal to the quality threshold, then a low-quality label is set for the image data corresponding to the time period.

[0011] Furthermore, the image upload module's sending strategy, in response to high-quality tags, includes sampling and sending the image data at preset intervals, including... Keyframe images are selected from the image data corresponding to the time period and sent according to the time interval; The time interval is adjusted based on the comprehensive defect score.

[0012] Furthermore, the cloud-based analysis module performs high-precision defect assessment based on the received image data, and conducts time-series analysis of the high-precision defect assessment results to provide early warnings of faults based on the changing trends of the high-precision defect assessment results. A high-precision defect assessment is performed on the received image data to obtain a recalculated comprehensive defect score. Time series data is established based on the recalculated comprehensive defect score, and trend analysis is performed on the time series data to identify quality anomaly patterns. Early warning of faults based on quality anomaly patterns.

[0013] Furthermore, the cloud-based analysis module performs high-precision defect assessment based on the received image data. The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several fine blocks. Based on the division of several fine blocks, the comprehensive defect score corresponding to the image data is calculated.

[0014] Furthermore, the cloud-based analysis module performs time-series analysis on the high-precision defect assessment results to provide early warnings of faults based on the changing trends of the high-precision defect assessment results. If the number of consecutive periods of rising comprehensive defect scores exceeds the quantity threshold, a quality trend fault warning will be issued. If the increase in the comprehensive defect score is greater than the sudden drop threshold in two adjacent time periods, an early warning of a sudden quality drop fault will be issued. The faults include quality trend faults and quality drop faults.

[0015] Compared with existing technologies, this invention acquires image data from the production line through an edge acquisition module at preset time intervals. An edge image processing module extracts contrast and surface texture features from the image data and identifies defect area contours based on standard product contours. An edge image analysis module calculates surface anomaly and defect anomaly indices, and a weighted comprehensive defect score is obtained. An edge upload module labels the image data according to the comprehensive defect score, enabling sampling transmission under high-quality labels and complete transmission under low-quality labels. A cloud analysis module performs high-precision defect assessment on the received image data and provides fault warnings through time-series analysis. This invention improves detection efficiency, reduces network bandwidth pressure through a tiered upload mechanism, and utilizes the high-precision assessment and time-series analysis capabilities of the cloud to determine quality degradation trends.

[0016] In particular, this invention considers a tiered upload mechanism. In practice, traditional quality inspection systems upload all collected data completely to the cloud, resulting in excessively large data streams and significantly increased network bandwidth pressure. This invention optimizes data transmission efficiency through a tag-based classification and tiered upload mechanism based on comprehensive defect scores. The system assigns high-quality or low-quality labels to image data for each time period based on the comprehensive defect score. When the score is below a quality threshold, only keyframe images are sampled and sent at preset intervals; when the score exceeds the threshold, all image data for that time period is sent in its entirety. This effectively reduces the transmission of redundant data and lowers network bandwidth usage while ensuring the complete collection of defect samples.

[0017] In particular, this invention considers the time-series analysis mechanism of edge cloud collaboration. In practice, traditional quality inspection methods are often limited to single-point defect identification, lacking effective means to determine the gradually accumulating process deviations and quality degradation trends during production. This invention achieves accurate early warning of quality trends by establishing a high-precision defect assessment and time-series analysis mechanism in the cloud. The system recalculates the comprehensive defect score based on a more refined block division standard for the received complete image data and constructs a time-series database. By setting quantity thresholds and sudden drop thresholds, the system can promptly warn of quality trend failures and quality drop failures, providing data support for continuous optimization of the production process.

[0018] In particular, this invention considers multi-dimensional feature analysis. In practice, a single detection index often fails to fully reflect the complex defect characteristics of a product surface, leading to insufficient detection accuracy. This invention achieves multi-dimensional evaluation of product quality by establishing a multi-dimensional analysis mechanism that combines surface features and contour features. The system simultaneously extracts contrast features, surface texture features, and defect region contour information at the edge, calculates the surface anomaly index and defect anomaly index respectively, and finally obtains a comprehensive defect score through weighted fusion. This multi-dimensional feature fusion analysis method improves the accuracy and reliability of defect identification, provides technical support for achieving high-precision detection, and reduces the false alarm rate. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the structural connection of an industrial internet quality inspection system based on edge cloud collaboration, as an embodiment of the invention. Figure 2 A logic block diagram for setting labels on image data corresponding to the time period in an embodiment of the invention; Figure 3 A logic block diagram for early warning of quality trend faults in an embodiment of the invention; Figure 4 This is a logic block diagram of an early warning system for sudden quality drops, as shown in an embodiment of the invention. Detailed Implementation

[0020] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0021] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0022] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0023] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0024] Please see Figure 1 As shown, Figure 1 This is a schematic diagram illustrating the structural connections of an edge-cloud collaborative industrial internet quality inspection system according to an embodiment of the invention. The edge-cloud collaborative industrial internet quality inspection system of the present invention includes: To achieve the above objectives, the present invention provides an industrial internet quality inspection system based on edge cloud collaboration, comprising: The edge acquisition module is used to acquire image data of the production line according to a preset time period; An edge image processing module, which is connected to the edge acquisition module, is used to extract contrast features and surface texture features from image data and identify the contour of defect areas based on a preset standard product contour. An edge image analysis module, which is connected to the edge image processing module, is used to calculate a surface anomaly index based on contrast features and surface texture features, calculate a defect anomaly index based on the defect region contour, and determine the weighted result of the surface anomaly index and the defect anomaly index as a comprehensive defect score. An edge upload module, connected to the edge image analysis, is used to upload image data based on a comprehensive defect score, including: Labels are assigned to the image data corresponding to the time period based on the comprehensive defect score; In response to a high-quality tag, the image data is sampled and sent at preset intervals; Alternatively, in response to a low-quality tag, send the complete image data; The cloud analysis module, which is connected to the edge upload module, is used to perform high-precision defect assessment based on the received image data, and to perform time-series analysis on the high-precision defect assessment results in order to provide early warning of faults based on the changing trend of the high-precision defect assessment results.

[0025] Specifically, the edge image processing module is used to identify the contour of defective regions based on a preset standard product contour, including, Identify the product outline of a target product in image data; The product outline is superimposed and compared with a preset standard product outline to identify the difference areas where the product outline and the standard outline differ. All discrepancy regions in the edge image are identified as the defect region contours.

[0026] Specifically, those skilled in the art need to predetermine the standard product profile. This can be generated by acquiring defect-free standard sample images, extracting and fitting edges, or directly exported from the product's CAD digital model. The constructed standard product profile will be stored as benchmark data on the edge side and in the cloud for profile comparison and defect identification during inspection.

[0027] Specifically, the edge image processing module overlays and compares the product outline with a preset standard product outline to identify the difference areas between the product outline and the standard outline, including: The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several blocks. Identify the areas where there are differences between the product outline and the standard product outline in the overlay comparison image, and filter out the difference blocks that meet the preset difference criteria; The difference blocks are connected and merged to form several closed difference regions.

[0028] In practice, blocks that meet the difference criteria are identified as having differences. The difference criteria are that the curvature difference ratio between the contour of the block and the contour of the corresponding standard product is greater than a predetermined curvature difference ratio threshold or the chromaticity difference ratio within the contour is greater than a predetermined chromaticity difference ratio threshold.

[0029] The curvature difference ratio threshold is selected within the range [0.15, 0.3], and the chromaticity difference ratio within the contour is selected within the range [0.25, 0.5].

[0030] Specifically, the edge image analysis module is used to calculate the surface anomaly index based on contrast features and surface texture features, including: The ratio of the contrast characteristic value to the contrast standard value is defined as the contrast index; The ratio of surface texture feature values ​​to texture standard values ​​is defined as the texture index; The surface anomaly index is obtained by weighted summation of the contrast index and the texture index.

[0031] Contrast feature value is the contrast of image data, surface texture feature value is the contrast of the corresponding gray-level co-occurrence matrix of image data. Image contrast reflects the overall brightness difference of the image, while surface texture feature value reflects the degree of messiness of the image. The contrast standard value and texture standard value are preset. Image data of standard parts are acquired, and the average contrast feature value is determined as the contrast standard value, and the average surface texture feature value is determined as the texture standard value.

[0032] The weights for both contrast index and texture index are 0.5.

[0033] Specifically, the standard contrast value and texture Specifically, the edge image analysis module is used to calculate a defect anomaly index based on the defect region contour, including: The ratio of the total area of ​​the defective region outline to the area of ​​the standard product is defined as the area index; The ratio of the number of differential regions in the defect region outline to the number of standard defects is determined as the quantity index; The defect anomaly index is obtained by weighted summation of the area index and the quantity index.

[0034] Specifically, the total area of ​​the defect region contour is calculated from the total number of pixels contained within the defect region contour.

[0035] Specifically, the standard product area refers to the theoretical total pixel area enclosed by a preset standard product outline under standard imaging conditions. The standard product area is a preset benchmark value, obtained directly by calculating the total number of pixels within the standard outline template, ensuring the uniformity and comparability of the area index calculation benchmark under different inspection batches and different equipment.

[0036] Specifically, the standard defect number is an empirical threshold pre-set by those skilled in the art based on process level and quality requirements. During stable operation of the production line, the average number of defective areas identified in a large number of product images judged as qualified is statistically analyzed, and this average value is determined as the standard defect number.

[0037] Specifically, the weighting coefficient for the area index is 0.6, and the weighting coefficient for the quantity index is 0.4. By giving the area index a higher weight, it is ensured that when a product has large-area defects, even if the number of defective areas is small, the overall defect score can still increase rapidly.

[0038] Specifically, the weighting coefficient for the surface anomaly index is 0.3, and the weighting coefficient for the defect anomaly index is 0.7. The system places greater emphasis on defects in the product's macroscopic structure and shape contour during evaluation, as contour defects have a more direct and severe impact on product quality.

[0039] Please see Figure 2 As shown, Figure 2 This is a logic block diagram illustrating how the edge upload module sets labels for image data corresponding to the time period according to an embodiment of the invention. The edge upload module sets labels for the image data corresponding to the time period based on a comprehensive defect score, including: If the overall defect score is less than the quality threshold, then a high-quality label is set for the image data corresponding to the time period. If the overall defect score is greater than or equal to the quality threshold, then a low-quality label is set for the image data corresponding to the time period.

[0040] Specifically, the quality threshold is preset by those skilled in the art. During stable operation of the production line, a large number of product images judged as qualified are statistically analyzed to obtain the average historical comprehensive defect score. The product of the average historical comprehensive defect score and the magnification factor is used as the quality threshold.

[0041] Specifically, the magnification factor ranges from [1.05, 1.15].

[0042] Specifically, the image upload module's sending strategy responds to high-quality tags, and includes sampling and sending the image data at preset intervals, including: Keyframe images are selected from the image data corresponding to the time period and sent according to the time interval; The time interval is adjusted based on the comprehensive defect score.

[0043] Specifically, the cloud-based analysis module performs high-precision defect assessment based on the received image data, performs time-series analysis of the high-precision defect assessment results, and provides early warnings of faults based on the changing trends of the high-precision defect assessment results. A high-precision defect assessment is performed on the received image data to obtain a recalculated comprehensive defect score. Time series data is established based on the recalculated comprehensive defect score, and trend analysis is performed on the time series data to identify quality anomaly patterns. Early warning of faults based on quality anomaly patterns.

[0044] Specifically, the cloud-based analysis module performs high-precision defect assessment based on the received image data. The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several fine blocks. Based on the division of several fine blocks, the comprehensive defect score corresponding to the image data is calculated.

[0045] Please see Figure 3 as well as Figure 4 As shown, Figure 3 This is a logic block diagram of an early warning quality trend fault according to an embodiment of the invention. Figure 4 This is a logic block diagram of an early warning system for sudden quality drops in an embodiment of the invention. The cloud analysis module performs time-series analysis of the high-precision defect assessment results to provide early warning of faults based on the changing trends of the high-precision defect assessment results. If the number of consecutive periods of rising comprehensive defect scores exceeds the quantity threshold, a quality trend fault warning will be issued. If the increase in the comprehensive defect score is greater than the sudden drop threshold in two adjacent time periods, an early warning of a sudden quality drop fault will be issued. The faults include quality trend faults and quality drop faults.

[0046] Specifically, the value range of the quantity threshold is [3,5].

[0047] Specifically, the threshold value ranges from [0.2, 0.4].

[0048] Specifically, there are no restrictions on the specific forms of the edge acquisition module, edge image processing module, edge image analysis module, edge upload module, and cloud analysis module. They can be composed of logical components, including field-programmable processors, computers, or microprocessors in computers, which will not be elaborated further.

[0049] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An industrial internet quality inspection system based on edge cloud collaboration, characterized in that, include: The edge acquisition module is used to acquire image data of the production line according to a preset time period; An edge image processing module, which is connected to the edge acquisition module, is used to extract contrast features and surface texture features from image data and identify the contour of defect areas based on a preset standard product contour. An edge image analysis module, which is connected to the edge image processing module, is used to calculate a surface anomaly index based on contrast features and surface texture features, calculate a defect anomaly index based on the defect region contour, and determine the weighted result of the surface anomaly index and the defect anomaly index as a comprehensive defect score. An edge upload module, connected to the edge image analysis, is used to upload image data based on a comprehensive defect score, including: Labels are assigned to the image data corresponding to the time period based on the comprehensive defect score; In response to a high-quality tag, the image data is sampled and sent at preset intervals; Alternatively, in response to a low-quality tag, send the complete image data; The cloud analysis module, which is connected to the edge upload module, is used to perform high-precision defect assessment based on the received image data, and to perform time-series analysis on the high-precision defect assessment results in order to provide early warning of faults based on the changing trend of the high-precision defect assessment results.

2. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The edge image processing module is used to identify the contour of defective regions based on a preset standard product contour, including: Identify the product outline of a target product in image data; The product outline is superimposed and compared with a preset standard product outline to identify the difference areas where the product outline and the standard outline differ. All discrepancy regions in the edge image are identified as the defect region contours.

3. The industrial internet quality inspection system based on edge cloud collaboration according to claim 2, characterized in that, The edge image processing module overlays and compares the product outline with a preset standard product outline to identify the difference areas between the product outline and the standard outline, including... The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several blocks. Identify the areas where there are differences between the product outline and the standard product outline in the overlay comparison image, and filter out the difference blocks that meet the preset difference criteria; The difference blocks are connected and merged to form several closed difference regions.

4. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The edge image analysis module is used to calculate the surface anomaly index based on contrast features and surface texture features, including: The ratio of the contrast characteristic value to the contrast standard value is defined as the contrast index; The ratio of surface texture feature values ​​to texture standard values ​​is defined as the texture index; The surface anomaly index is obtained by weighted summation of the contrast index and the texture index.

5. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The edge image analysis module is used to calculate the defect anomaly index based on the defect region contour, including: The ratio of the total area of ​​the defective region outline to the area of ​​the standard product is defined as the area index; The ratio of the number of differential regions in the defect region outline to the number of standard defects is determined as the quantity index; The defect anomaly index is obtained by weighted summation of the area index and the quantity index.

6. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The edge upload module sets labels for the image data corresponding to the time period based on a comprehensive defect score, including... If the overall defect score is less than the quality threshold, then a high-quality label is set for the image data corresponding to the time period. If the overall defect score is greater than or equal to the quality threshold, then a low-quality label is set for the image data corresponding to the time period.

7. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The image upload module's sending strategy responds to high-quality tags, and includes sampling and sending the image data at preset intervals, including... Keyframe images are selected from the image data corresponding to the time period and sent according to the time interval; The time interval is adjusted based on the comprehensive defect score.

8. The industrial internet quality inspection system based on edge cloud collaboration according to claim 1, characterized in that, The cloud-based analysis module performs high-precision defect assessment based on the received image data, and performs time-series analysis on the high-precision defect assessment results to provide early warnings of faults based on the changing trends of the high-precision defect assessment results. A high-precision defect assessment is performed on the received image data to obtain a recalculated comprehensive defect score. Time series data is established based on the recalculated comprehensive defect score, and trend analysis is performed on the time series data to identify quality anomaly patterns. Early warning of faults based on quality anomaly patterns.

9. The industrial internet quality inspection system based on edge cloud collaboration according to claim 8, characterized in that, The cloud-based analysis module performs high-precision defect assessment based on the received image data. The product outline is overlaid with the standard product outline to form an overlay comparison image, which is then divided into several fine blocks. Based on the division of several fine blocks, the comprehensive defect score corresponding to the image data is calculated.

10. The industrial internet quality inspection system based on edge cloud collaboration according to claim 8, characterized in that, The cloud-based analysis module performs time-series analysis on the high-precision defect assessment results to provide early warnings of faults based on the changing trends of the high-precision defect assessment results. If the number of consecutive periods of rising comprehensive defect scores exceeds the quantity threshold, a quality trend fault warning will be issued. If the increase in the comprehensive defect score is greater than the sudden drop threshold in two adjacent time periods, an early warning of a sudden quality drop fault will be issued. The faults include quality trend faults and quality drop faults.

Images