Edge computing method and system for finished product inspection based on AI vision
By adopting AI vision's manufacturing finished product detection edge computing method in the manufacturing industry, using edge computing and lossless compression technology to perform image recognition and detection on local devices, the problems of data transmission noise and resource waste in complex environments are solved, and efficient Safe inspection of finished products has improved the production efficiency of manufacturing enterprises.
Patent Information
- Application Number
- CN202410502683.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-04-24
AI Technical Summary
Existing edge computing technologies cannot meet the noise problems during data transmission in complex manufacturing environments, and the identification prediction of finished manufacturing requires a large amount of data training models, resulting in waste of resources and time consumption.
Using AI vision-based manufacturing edge computing method, image data is collected through imaging equipment for pre-processing, image recognition and detection are used on local computer equipment, AI vision model is established, and lossless compression technology and delay compensation are used. Optimize the communication channel, select the optimal channel for data transmission, identify defective products and distinguish normal products in real time, and feedback optimize the model to improve detection efficiency and accuracy.
Real-time operation of AI models in the manufacturing industry, reduces data transmission delay and network dependence, protects data privacy and security, improves the real-time and security of finished product inspection, and brings more efficient production to manufacturing enterprises. Way.
Smart Images

Figure CN118333994B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of edge computing and visual recognition technology, and specifically to an edge computing method and system for manufacturing finished product detection based on AI vision. Background Art
[0002] With the continuous development of artificial intelligence technology, AI vision is increasingly used in the manufacturing industry. Quality inspection of finished products is a crucial part of the production process, and the use of AI vision technology for finished product inspection can effectively improve production efficiency and product quality. However, traditional cloud-based AI vision inspection systems have problems such as data transmission delays and privacy security. Therefore, the method of deploying AI models on the device side using edge computing has gradually attracted attention.
[0003] For example, the Chinese patent with publication number CN116822319A discloses an active prediction and control method for industrial product manufacturing quality based on cloud-edge collaborative computing, which includes the following steps: collecting a large amount of data related to the quality of complex industrial products during the manufacturing process through the device perception layer; reliably deploying the important quality-related parameter selection rules and the quality active prediction model to the edge computing layer; the edge computing layer uses the important quality-related parameter selection rules to screen the transmitted real-time production data of industrial products, calculate the difference between the qualified rate of actual output products and the qualified rate of industrial products containing the predicted results; and record the industrial product quality data obtained by the detection. Compared with the prior art, the advantages of this invention are: real-time feedback of industrial product manufacturing quality prediction results and issuance of corresponding processing technology adjustment control instructions, which well copes with the deviation of quality prediction results.
[0004] The defects of the above patent are: 1) When collecting finished products for transmission, the existing edge computing technology cannot meet the needs of complex environments (manufactured product production workshops) and there is a lot of noise; 2) To identify and predict finished products, a large amount of data area training models is required, which wastes a lot of resources and time when processing simple manufactured products. Summary of the invention
[0005] In view of the shortcomings of the prior art, the present invention proposes an edge computing method and system for finished product detection based on AI vision, which uses a camera device to collect image data of finished products in a production environment, pre-processes the collected image data of finished products, and uses an edge computing strategy to transmit the pre-processed image data of finished products to a local computer device for image recognition and detection. An AI vision model is established to identify and detect the image data of finished products transmitted to the local computer device, identify defective products and distinguish normal products, and process them. The model is optimized based on feedback from the monitoring results, and the AI vision model and edge computing method are continuously optimized to improve the efficiency and accuracy of finished product detection.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] The edge computing method for manufacturing finished product detection based on AI vision includes the following specific steps:
[0008] Step S1: using a camera to collect image data of finished products in a production environment;
[0009] Step S2: pre-processing the collected finished product image data;
[0010] Step S3: using edge computing strategies, the pre-processed image data of the manufactured product is transmitted to a local computer device for image recognition and detection;
[0011] Step S4: Establish an AI visual model to identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them;
[0012] Step S5: Feedback and optimize the model based on the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product inspection.
[0013] Specifically, the preprocessing in step S2 includes removing noise, including noise in the device and the image.
[0014] Specifically, the specific steps of step S3 are:
[0015] Step S301: compressing the pre-processed manufacturing product data using lossless compression technology;
[0016] Step S302: determine the edge nodes, configure the determined edge nodes, deploy edge computing resources, transmit the compressed finished product data, compensate for the communication delay during the transmission process, and select the communication channel;
[0017] Step S303: Calculate the delay of the local computer device, the calculation formula is: Among them, T bs represents the delay of the local computer device, t s Indicates the time to complete the calculation of data volume s, js bs Indicates the computing power of the local computer device, which compensates for the communication delay during the transmission process. The calculation formula is:
[0018] T bc =T bs +(t fs -t bz ),
[0019] Among them, Tbc represents the compensation of communication delay during transmission, t fs Indicates the time it takes for the camera device to send image data to the local computer device, t bz Indicates the standard time for the camera device to send image data to the local computer device;
[0020] Step S304: Set the communication channel set for transmitting the compressed finished product image data to L, where L = {l1, l2, ..., l n}, calculate the lth n The priority of a communication channel is calculated as follows:
[0021]
[0022] in, Indicates the first i The priority of a communication channel, eff represents the signal quality indicator, α represents the capacity adjustment factor, r represents the historical transmission rate, e represents the exponential function, sjl represents the data volume of the compressed finished product image data, and Z fs Indicates the period during which the camera device sends image data to the local computer device, t min It represents the minimum time required for the local computer device to process the image data received per unit cycle, μ represents the penalty factor, Indicates the first i The load of each communication channel, Indicates the first i The maximum load of a communication channel, max(·) represents the maximum value function, The corresponding communication channel is used as a communication channel for the camera device to send image data to the local computer device.
[0023] Specifically, the lossless compression technology in step S301 includes: Huffman algorithm, LZ77 algorithm and LZR algorithm.
[0024] Specifically, the specific steps of step S4 are:
[0025] Step S401: decompressing the received compressed finished product image data to obtain decompressed finished product image data;
[0026] Step S402: preprocessing the decompressed finished product image data, the preprocessing including: image segmentation and image denoising, image denoising is to remove the noise in the characteristic area of each finished product image, image segmentation is to segment each denoised finished product image and remove the background image outside the characteristic area of each finished product image;
[0027] Step S403: Set the pre-processed finished product image set as ZZ, where ZZ = {zz1, zz2, ..., zz m}, where zz m represents the mth finished product image after preprocessing, and obtains the area and mass of the finished product after preprocessing in the target image during the production process. The sets of the area and mass of the finished product after preprocessing are set as MJ and ZL respectively, MJ = {mj1, mj2, ..., mj m}, ZL={zl1, zl2,...,zl m}, mj m represents the area of the mth manufactured product after preprocessing, zl m represents the quality of the mth manufactured product after preprocessing;
[0028] Step S404: Calculate the area ratio and mass ratio of the i-th manufactured product after preprocessing, and the calculation formula is:
[0029]
[0030] in, represents the area ratio of the i-th manufactured product after preprocessing, represents the mass ratio of the i-th manufactured product after pretreatment, mj i represents the area of the i-th manufactured product after preprocessing, zl i represents the quality of the i-th manufactured product after pretreatment, mj bz Indicates the standard area of the finished product, zl bz Indicates the standard quality of manufactured products;
[0031] Step S405: Establish an AI visual model to evaluate the quality of the i-th manufactured product after preprocessing. The specific evaluation formula is:
[0032]
[0033] Among them, PG i represents the quality of the i-th manufactured product, λ i represents the correction factor for the evaluation of the i-th manufactured product after preprocessing;
[0034] Step S406: According to the quality evaluation value of the i-th manufactured product after preprocessing, defective products and normal products are identified and processed.
[0035] Specifically, the specific method of step S406 is: set the quality threshold τ1, the area ratio threshold τ2 and the quality ratio threshold τ3, and when the i-th manufactured product is greater than the quality threshold τ1, the area ratio threshold τ2 and the quality ratio threshold τ3, it is determined to be a normal product.i When it is less than τ1, it is judged as defective. i is greater than τ1, and when Less than τ2 or When it is less than τ3, it is judged as a defective product.
[0036] The edge computing system for finished product inspection based on AI vision includes: image data acquisition module, image data preprocessing module, edge computing module, recognition and detection module and feedback optimization module;
[0037] The image data acquisition module is used to collect image data of finished products manufactured in a production environment using a camera device;
[0038] The image data preprocessing module is used to preprocess the collected finished product image data;
[0039] The edge computing module is used to transmit the pre-processed finished product image data to the local computer device for image recognition and detection using the edge computing strategy;
[0040] The identification and detection module is used to establish an AI visual model, identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them;
[0041] The feedback optimization module is used to perform feedback optimization on the model according to the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product detection.
[0042] Specifically, the edge computing module includes a delay compensation unit and an optimal channel selection unit;
[0043] The delay compensation unit is used to compensate for the communication delay during the transmission process between the camera device and the local computer device;
[0044] The optimal channel selection unit is used to select the optimal communication channel during the transmission process between the camera device and the local computer device.
[0045] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of an edge computing method for manufacturing finished product detection based on AI vision are implemented.
[0046] A computer-readable storage medium stores computer instructions, which, when executed, execute the steps of an edge computing method for manufacturing finished product detection based on AI vision.
[0047] Compared with the prior art, the present invention has the following beneficial effects:
[0048] 1. The present invention proposes an edge computing method for finished product inspection based on AI vision. Edge computing processes data on the device side and selects the optimal communication channel in the network transmission between the camera device and the local computer device, so that the AI model can run in real time, reducing data transmission delay and dependence on network connection, while protecting the privacy and security of the data.
[0049] 2. The present invention proposes an edge computing method for finished product detection based on AI vision. In the manufacturing industry, the edge computing method for finished product detection based on AI vision has better real-time performance and security, bringing a more efficient production method to manufacturing enterprises. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] Figure 1 This is a flow chart of the edge computing method for finished product detection based on AI vision of the present invention;
[0051] Figure 2 This is an edge computing detection flow chart of the edge computing method for manufacturing finished product detection based on AI vision of the present invention;
[0052] Figure 3 This is an architecture diagram of the edge computing system for finished product detection based on AI vision of the present invention;
[0053] Figure 4 This is a diagram of an electronic device for the edge computing method for finished product detection based on AI vision of the present invention. DETAILED DESCRIPTION
[0054] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, in the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate the orientation or position relationship based on the orientation or position relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "No. 1", "No. 2" and "No. 3" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. The present invention is further explained below in conjunction with specific embodiments.
[0055] Example 1
[0056] See also Figure 1-Figure 2 , an embodiment provided by the present invention: an edge computing method for manufacturing finished product detection based on AI vision, comprising the following specific steps:
[0057] Step S1: using a camera to collect image data of finished products in a production environment;
[0058] Step S2: pre-processing the collected finished product image data;
[0059] Step S3: using edge computing strategies, the pre-processed image data of the manufactured product is transmitted to a local computer device for image recognition and detection;
[0060] Step S4: Establish an AI visual model to identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them;
[0061] Step S5: Feedback and optimize the model based on the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product inspection.
[0062] The preprocessing in step S2 includes removing noise, including noise in the device and the image.
[0063] The specific steps of step S3 are:
[0064] Step S301: compressing the pre-processed manufacturing product data using lossless compression technology;
[0065] Step S302: determine the edge nodes, configure the determined edge nodes, deploy edge computing resources, transmit the compressed finished product data, compensate for the communication delay during the transmission process, and select the communication channel;
[0066] Step S303: Calculate the delay of the local computer device, the calculation formula is: Among them, T bs represents the delay of the local computer device, t s Indicates the time to complete the calculation of data volume s, js bs Indicates the computing power of the local computer device, which compensates for the communication delay during the transmission process. The calculation formula is:
[0067] T bc =T bs +(t fs -t bz ),
[0068] Among them, T bc represents the compensation of communication delay during transmission, t fs Indicates the time it takes for the camera device to send image data to the local computer device, t bz Indicates the standard time for the camera device to send image data to the local computer device;
[0069] Step S304: Set the communication channel set for transmitting the compressed finished product image data to L, where L = {l1, l2, ..., ln}, calculate the lth n The priority of a communication channel is calculated as follows:
[0070]
[0071] in, Indicates the first i The priority of a communication channel, eff represents the signal quality indicator, α represents the capacity adjustment factor, r represents the historical transmission rate, e represents the exponential function, sjl represents the data volume of the compressed finished product image data, and Z fs Indicates the period during which the camera device sends image data to the local computer device, t min It represents the minimum time required for the local computer device to process the image data received per unit cycle, μ represents the penalty factor, Indicates the first i The load of each communication channel, Indicates the first i The maximum load of a communication channel, max(·) represents the maximum value function, The corresponding communication channel is used as a communication channel for the camera device to send image data to the local computer device.
[0072] The lossless compression techniques in step S301 include: Huffman algorithm, LZ77 algorithm and LZR algorithm.
[0073] The Huffman algorithm is a compression algorithm based on the frequency of character occurrence. It reduces the file size by constructing an optimal coding tree and using shorter codes to represent characters with higher frequencies. The LZ77 algorithm is the basis of the LZW algorithm. They use sliding windows and dictionary technology for compression.
[0074] The specific steps of step S4 are:
[0075] Step S401: decompressing the received compressed finished product image data to obtain decompressed finished product image data;
[0076] Step S402: preprocessing the decompressed finished product image data, the preprocessing including: image segmentation and image denoising, image denoising is to remove the noise in the characteristic area of each finished product image, image segmentation is to segment each denoised finished product image and remove the background image outside the characteristic area of each finished product image;
[0077] The preprocessing here is different from the preprocessing in step S2. The preprocessing in step S2 is to remove the particle noise, equipment noise and Gaussian noise in the image, while the preprocessing in step S402 is to remove the background image outside the feature area of the finished product;
[0078] Step S403: Set the pre-processed finished product image set as ZZ, where ZZ = {zz1, zz2, ..., zz m}, where zz m represents the mth finished product image after preprocessing, and obtains the area and mass of the finished product after preprocessing in the target image during the production process. The sets of the area and mass of the finished product after preprocessing are set as MJ and ZL respectively, MJ = {mj1, mj2, ..., mj m}, ZL={zl1, zl2,...,zl m}, mj m represents the area of the mth manufactured product after preprocessing, zl m represents the quality of the mth manufactured product after preprocessing;
[0079] Step S404: Calculate the area ratio and mass ratio of the i-th manufactured product after preprocessing, and the calculation formula is:
[0080]
[0081] in, represents the area ratio of the i-th manufactured product after preprocessing, represents the mass ratio of the i-th manufactured product after pretreatment, mj i represents the area of the i-th manufactured product after preprocessing, zl i represents the quality of the i-th manufactured product after pretreatment, mj bz Indicates the standard area of the finished product, zl bz Indicates the standard quality of manufactured products;
[0082] Step S405: Establish an AI visual model to evaluate the quality of the i-th manufactured product after preprocessing. The specific evaluation formula is:
[0083]
[0084] Among them, PG i represents the quality of the i-th manufactured product, λ i represents the correction factor for the evaluation of the i-th manufactured product after preprocessing;
[0085] Step S406: According to the quality evaluation value of the i-th manufactured product after preprocessing, identify defective products and distinguish normal products, set the quality threshold τ1, area ratio threshold τ2 and quality ratio threshold τ3, and when the i-th manufactured product is greater than the quality threshold τ1, area ratio threshold τ2 and quality ratio threshold τ3, it is judged as a normal product. When PG i When it is less than τ1, it is judged as defective. iis greater than τ1, and when Less than τ2 or When it is less than τ3, it is judged as a defective product and processed.
[0086] Example 2
[0087] See also Figure 3 , another embodiment provided by the present invention: an edge computing system for manufacturing finished product detection based on AI vision, comprising: an image data acquisition module, an image data preprocessing module, an edge computing module, a recognition and detection module and a feedback optimization module;
[0088] The image data acquisition module is used to collect image data of finished products manufactured in a production environment using a camera device;
[0089] The image data preprocessing module is used to preprocess the collected finished product image data;
[0090] The edge computing module is used to transmit the pre-processed finished product image data to the local computer device for image recognition and detection using the edge computing strategy;
[0091] The identification and detection module is used to establish an AI visual model, identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them;
[0092] The feedback optimization module is used to perform feedback optimization on the model according to the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product detection.
[0093] The edge computing module includes a delay compensation unit and an optimal channel selection unit;
[0094] The delay compensation unit is used to compensate for the communication delay during the transmission process between the camera device and the local computer device;
[0095] The optimal channel selection unit is used to select the optimal communication channel during the transmission process between the camera device and the local computer device.
[0096] Example 3
[0097] See also Figure 4 , an electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of an edge computing method for manufacturing finished product detection based on AI vision are implemented.
[0098] A computer-readable storage medium stores computer instructions, which, when executed, execute the steps of an edge computing method for manufacturing finished product detection based on AI vision.
[0099] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0100] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0101] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0102] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0103] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.
[0104] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. The edge computing method for finished product detection based on AI vision is characterized by: The specific steps include: Step S1: using a camera to collect image data of finished products in a production environment; Step S2: pre-processing the collected finished product image data; Step S3: using edge computing strategies, the pre-processed image data of the manufactured product is transmitted to a local computer device for image recognition and detection; Step S4: Establish an AI visual model to identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them; Step S5: Feedback and optimize the model based on the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product inspection; The specific steps of step S3 are: Step S301: compressing the pre-processed manufacturing product data using lossless compression technology; Step S302: determine the edge nodes, configure the determined edge nodes, deploy edge computing resources, transmit the compressed finished product data, compensate for the communication delay during the transmission process, and select the communication channel; Step S303: Calculate the delay of the local computer device, the calculation formula is: Among them, T bs represents the delay of the local computer device, t s Indicates the time to complete the calculation of data volume s, js bs Indicates the computing power of the local computer device, which compensates for the communication delay during the transmission process. The calculation formula is: T bc =T bs +(t fs -t bz ), Among them, T bc represents the compensation of communication delay during transmission, t fs Indicates the time it takes for the camera device to send image data to the local computer device, t bz Indicates the standard time for the camera device to send image data to the local computer device; Step S304: Set the communication channel set for transmitting the compressed finished product image data to L, where L = {l1, l2, ..., l n }, calculate the lth n The priority of a communication channel is calculated as follows: in, Indicates the first i The priority of a communication channel, eff represents the signal quality indicator, α represents the capacity adjustment factor, r represents the historical transmission rate, e represents the exponential function, sjl represents the data volume of the compressed finished product image data, and Z fs Indicates the period during which the camera device sends image data to the local computer device, t min It represents the minimum time required for the local computer device to process the image data received per unit cycle, μ represents the penalty factor, Indicates the first i The load of each communication channel, Indicates the first i The maximum load of a communication channel, max(·) represents the maximum value function, The corresponding communication channel is used as a communication channel for the camera device to send image data to the local computer device; The lossless compression techniques in step S301 include: Huffman algorithm, LZ77 algorithm and LZR algorithm; The specific steps of step S4 are: Step S401: decompressing the received compressed finished product image data to obtain decompressed finished product image data; Step S402: preprocessing the decompressed finished product image data, the preprocessing including: image segmentation and image denoising, image denoising is to remove the noise in the characteristic area of each finished product image, image segmentation is to segment each denoised finished product image and remove the background image outside the characteristic area of each finished product image; Step S403: Set the pre-processed finished product image set as ZZ, ZZ = {zz1, zz2, ..., zz m }, where zz m represents the mth finished product image after preprocessing, and obtains the area and mass of the finished product after preprocessing in the target image during the production process. The sets of the area and mass of the finished product after preprocessing are set as MJ and ZL respectively, MJ = {mj1, mj2, ..., mj m }, ZL={zl1,zl2,...,zl m }, mj m represents the area of the mth manufactured product after preprocessing, zl m represents the quality of the mth manufactured product after preprocessing; Step S404: Calculate the area ratio and mass ratio of the i-th manufactured product after preprocessing, and the calculation formula is: in, represents the area ratio of the i-th manufactured product after preprocessing, represents the mass ratio of the i-th manufactured product after pretreatment, mj i represents the area of the i-th manufactured product after preprocessing, zl i represents the quality of the i-th manufactured product after pretreatment, mj bz Indicates the standard area of the finished product, zl bz Indicates the standard quality of manufactured products; Step S405: Establish an AI visual model to evaluate the quality of the i-th manufactured product after preprocessing. The specific evaluation formula is: Among them, PG i represents the quality of the i-th manufactured product, λ i represents the correction factor for the evaluation of the i-th manufactured product after preprocessing; Step S406: According to the quality evaluation value of the i-th manufactured product after preprocessing, defective products and normal products are identified and processed.
2. The edge computing method for manufacturing finished product detection based on AI vision as claimed in claim 1, characterized in that: The preprocessing in step S2 includes removing noise, including noise in the device and the image.
3. The edge computing method for finished product detection based on AI vision as claimed in claim 1, characterized in that: The specific method of step S406 is: set the quality threshold τ1, the area ratio threshold τ2 and the quality ratio threshold τ3, and when the i-th manufactured product is greater than the quality threshold τ1, the area ratio threshold τ2 and the quality ratio threshold τ3, it is judged as a normal product. i When it is less than τ1, it is judged as defective. i is greater than τ1, and when Less than τ2 or When it is less than τ3, it is judged as a defective product.
4. An edge computing system for finished product detection based on AI vision, which is implemented based on the edge computing method for finished product detection based on AI vision according to any one of claims 1 to 3, characterized in that: include: Image data acquisition module, image data preprocessing module, edge computing module, recognition and detection module and feedback optimization module; The image data acquisition module is used to collect image data of finished products manufactured in a production environment using a camera device; The image data preprocessing module is used to preprocess the collected finished product image data; The edge computing module is used to transmit the pre-processed finished product image data to the local computer device for image recognition and detection using the edge computing strategy; The identification and detection module is used to establish an AI visual model, identify and detect the finished product image data transmitted to the local computer device, identify defective products and distinguish normal products, and process them; The feedback optimization module is used to perform feedback optimization on the model according to the monitoring results, continuously optimize the AI vision model and edge computing method, and improve the efficiency and accuracy of finished product detection.
5. The AI vision-based finished product detection edge computing system according to claim 4, characterized in that: The edge computing module includes a delay compensation unit and an optimal channel selection unit; The delay compensation unit is used to compensate for the communication delay during the transmission process between the camera device and the local computer device; The optimal channel selection unit is used to select the optimal communication channel during the transmission process between the camera device and the local computer device.
6. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the edge computing method for finished product detection based on AI vision described in any one of claims 1-3 are implemented.
7. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the steps of the edge computing method for finished product detection based on AI vision described in any one of claims 1-3 are executed.
