A pure structural consumables automatic generation assembly traceability system and method

Abstract

The present invention discloses a pure structural consumables automatic generation assembly traceability system and method, which relates to the technical field of product identification traceability, and the method comprises the steps of: S1, receiving an order and assigning a component code; S2, printing a code on a component; S3, performing quality inspection and marking on the component, and transmitting back the code and quality inspection results; S4, querying the component quality inspection results, ensuring that all components are qualified before assembly, and binding the assembled finished product with the component code; S5, binding the finished product code with the finished product packaging code; S6, analyzing and repairing the unqualified components in the quality inspection, and re-entering step S4 after the repaired components are qualified in the quality inspection; the components that cannot be repaired are classified as unqualified products; S7, completing the traceability of finished products and components through coding information. The present invention traces back to the production process of components through coding information; and identifies component defects before assembling the finished product to improve the qualified rate of finished products.

Description

Technical Field

[0001] The present invention relates to the technical field of product identification and traceability, and in particular to a system and method for automatically generating and assembling traceability for pure structural consumables. Background Art

[0002] In the current field of medical pelvic floor screening equipment manufacturing, the widely used pure structural consumables on the market are composed of a variety of sub-structure modules. The existing production process usually directly assembles the sub-structures after the modular production is completed, and lacks quality inspection steps for each link in the production process.

[0003] This practice results in quality problems often being discovered only in the final assembly stage, causing relatively lagging production process management and the defective rate of the entire machine being equal to the sum of the defective rates of all components, resulting in a huge waste of resources.

[0004] Due to the lack of intermediate inspection links, factories have to plan actual production volume based on the number of customer orders and taking into account the maximum possible loss rate to ensure that delivery requirements can be met, which not only increases production costs but also reduces production efficiency and flexibility. In addition, traditional quality control methods rely on manual operations or simple automated testing, which have problems such as low efficiency, prone to errors, and difficulty in achieving transparent management throughout the process, and cannot fully adapt to the development requirements of intelligent manufacturing. Summary of the invention

[0005] The purpose of the present invention is to provide a purely structural consumables automatic generation and assembly traceability system and method to solve the problems raised in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for automatically generating assembly traceability of pure structural consumables, the method comprising the steps of: S1. After receiving the order, the matching table of finished product and component quality inspection results is captured from the finished product BOM, and the component code is assigned to the order according to the coding rules; S2. The laser engraving printer obtains the component code and prints the code on the designated area of ​​the component; S3, conduct real-time quality inspection and marking of parts, and transmit the part code and quality inspection results back; S4. When assembling parts, use QR code scanning or image recognition to collect the codes on the surface of the parts, query the parts quality inspection results, ensure that all parts are qualified before assembly, and rebind the reassembled finished product with the parts code; S5. Each finished product has an independent package. The finished product packaging code corresponds to the finished product one by one, and the finished product code and the finished product packaging code are bound; S6, analyzing the parts that failed the quality inspection to determine whether they can be repaired; if they can be repaired, the repaired parts will be re-inspected for quality. If the quality inspection is qualified, the quality inspection mark of the parts will be updated and the process will proceed to step S4; S7. Complete the traceability of finished products and parts by binding the finished product packaging code with the finished product code.

[0007] According to the above scheme, the component code grabs the matching table of finished product and component quality inspection from the finished product bill of materials, assigns the component code to the order according to the customized coding rules, and generates an order matching table; the order matching table includes the finished product code, component code and quality inspection status; The component code, each component has a unique component code; Use a laser engraving printer to print the part number in the designated area of ​​the part.

[0008] According to the above scheme, the quality detection includes: Collect high-definition images of parts after printing and coding, and perform denoising, image enhancement and grayscale conversion on the collected high-definition images of parts; The denoising method uses a Gaussian filter to remove noise points in the image and reduce interference. The formula is as follows: ; Among them, G(x,y) represents the denoised image; w(i,j) represents the Gaussian weight; I(x,y) represents the original image; k represents the window radius of the filter; The image enhancement improves the quality of the image by contrast enhancement, and the formula is as follows: ; Among them, H(x,y) represents the enhanced image; μ represents the mean of the image; σ represents the standard deviation; The grayscale conversion uses a weighted average method to calculate the grayscale value of each pixel and convert the color image into a grayscale image; The processed image is input into the component quality inspection model, and the component quality inspection model is used to determine whether the component is qualified; components that pass the quality inspection are marked as "PASS", and components that fail the quality inspection are marked as "NG".

[0009] According to the above scheme, the component quality inspection model includes a convolutional layer, a pooling layer and a fully connected layer; The convolution layer uses convolution kernels of different sizes to perform convolution operations on the image and extract local features. After the convolution operation is completed, the ReLU activation function is used to introduce nonlinearity so that the model can learn more complex features. The formula is as follows: F(X)=max(0,X); Among them, F(X) represents the output value of the activation function; X represents the input value of the activation function; The pooling layer, after the convolutional layer, is used to reduce the size of the feature map, reduce the amount of calculation, and retain key information; The fully connected layer, after the convolutional layer and the pooling layer, integrates the extracted features to form the output. The formula is as follows: Y = λ(W × X + b); Among them, Y represents the output value; λ represents the ReLU activation function; W represents the weight matrix; b represents the bias term; Use the Softmax activation function to convert the output into two categories: "qualified" and "unqualified". The formula is as follows: ; Among them, Y i Expressed as the predicted probability of category i; z i is represented as the original output of category i; C is represented as the total number of categories; j is represented as the index of the category.

[0010] According to the above scheme, during the training process, the component quality inspection model calculates the predicted value of each sample through forward propagation, compares it with the true label, and calculates the error through the loss function. The loss function formula is as follows: ; Among them, L represents the loss function; N represents the total number of samples; y i Represented as the true label of the i-th sample; represents the probability predicted by the component quality inspection model; The gradient of the loss relative to the network parameters is calculated by back propagation, and the optimizer is used to update the parameters of the parts quality inspection model. The formula is as follows: ; Among them, θ t Represents the parameters of the current component quality inspection model; η represents the learning rate, which controls the step size of parameter update; m t Expressed as the mean of the gradient; v t It is represented as the variance of the gradient; ε is represented as a constant to prevent the denominator from being zero; The parameters of the component quality inspection model are updated once in each iteration until the component quality inspection model converges.

[0011] According to the above scheme, before the parts are ready for assembly, an image recognition device is used to read the unique coding information on the surface of each part and query the quality inspection results of the parts; when the quality inspection results of each part are qualified, that is, marked as "PASS", the assembly process will be entered; when the quality inspection results of any part are unqualified, that is, marked as "NG", the assembly process cannot be entered; After all parts and components are qualified in quality inspection and assembled into finished products, the assembled finished products will be assigned a unique finished product code, and the finished product code will be bound to the parts and components code.

[0012] According to the above scheme, the parts that fail the quality inspection are analyzed and repaired; the repaired parts are re-inspected for quality. If the quality inspection is qualified, the mark of the parts is changed from "NG" to "PASS" and the process goes to step S4; if the parts cannot be repaired, the parts are classified as defective and cannot be assembled; For structural parts, if the quality inspection fails, it is necessary to inspect whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the failure of the quality inspection of the parts is due to appearance defects, the maintenance personnel will clean and repair the parts. After cleaning and repair, the quality inspection will be carried out again. If the quality inspection is qualified, the process proceeds to step S4, otherwise it is classified as a defective product and cannot be assembled; when the failure of the quality inspection of the parts is due to tolerances that do not meet the requirements, the tolerance data of the parts must be recorded and they will be classified as defective products and cannot be assembled.

[0013] According to the above solution, before the finished product code and the finished product packaging code are bound, the finished product code and the finished product packaging code need to be printed; The printing device obtains the finished product code and the packaging code, creates a timer, and sends the finished product code to the inkjet printer; after the inkjet printer receives the finished product code, when sensing the electrode signal, it starts printing the finished product code and sends a "start printing" signal to the printing device; after the inkjet printer completes printing, it returns the printed content to the printing device; After receiving the print content, the printing device turns off the timer and adds the packaging code to the print list; during the printing process of the inkjet printer, if the timer times out, a printing abnormality alarm is triggered; When the printing device receives the print completion signal from the cutter printer, the cutter printer will feed back the print content to the printing device; the printing device sends the packaging code to the cutter printer, and receives a successful reception confirmation message from the cutter printer; When the number of packaging codes in the print list reaches or exceeds m, where m is a positive integer set by the system; and after the sending and receiving operations of all print tasks have been completed, the printing device will remove the processed packaging code items from the queue.

[0014] According to the above solution, after the finished product is packaged, the packaging code and the box code are bound and delivered to the client; Through the coding information of parts and finished products, the entire production and quality inspection process can be traced back to the source, and the entire process of parts production, assembly, inspection and repair can be traced; through the coding information of parts, the delivery, sales and use process of the finished product can be reversely traced.

[0015] A pure structural consumables automatic generation assembly traceability system, the system includes an order receiving module, a component code printing module, a component quality inspection module, a finished product assembly and coding binding module, a packaging code and finished product code binding module, a non-conforming component analysis module and a traceability module; The order receiving module is used to receive the customer's order and extract the quality inspection matching table of finished products and parts from the finished product bill of materials; assign a part code to each order according to the customer's customized coding rules and generate an order matching table; the order matching table includes the finished product code, part code and quality inspection status; The component code printing module assigns a unique code to each component and sends the code of each component to the laser engraving printer, which prints the component code on a designated area on the surface of the component; The component quality inspection module collects high-definition images of components after printing and coding, performs denoising, image enhancement and grayscale conversion on the images, and then inputs them into the quality inspection model for judgment to determine whether the component quality inspection is qualified; components that pass the quality inspection are marked as "PASS" and components that fail the quality inspection are marked as "NG"; The finished product assembly and coding binding module uses an image recognition device to read the unique coding information on the surface of each component and query the component quality inspection results; after all components have passed the quality inspection, the finished product is assembled, the components that have passed the quality inspection are assembled, and a unique finished product code is assigned to the assembled finished product; after the assembly is completed, the finished product code is bound to the component code to ensure the transparency and traceability of the assembly process; the finished product assembly and coding binding module also includes finished product code and packaging code generation, finished product code printing function and packaging code printing function; The unqualified parts analysis module is used to conduct a detailed analysis of the parts that fail the quality inspection and determine the type of problem of the unqualified quality inspection; for repairable parts, they are repaired and re-inspected. If the quality inspection result is qualified, the mark of the part is changed from "NG" to "PASS", and the finished product assembly and coding binding module is re-entered; if it cannot be repaired, it is classified as a defective product and cannot enter the finished product assembly and coding binding module; for structural parts, if the quality inspection fails, it is necessary to check whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the unqualified quality inspection of the parts is due to appearance defects, the maintenance personnel will clean and repair the parts. After the cleaning and repair are completed, the quality inspection will be carried out again. If the quality inspection is qualified, the finished product assembly and coding binding module will be re-entered, otherwise it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; when the unqualified quality inspection of the parts is due to the tolerance not meeting the requirements, the tolerance data of the parts needs to be recorded, and it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; The traceability module records and stores the coding, production, assembly, testing, repair, delivery and sales information of parts and finished products; through the coding information of parts and finished products, the traceability of the entire production, quality inspection and sales process is completed, and the entire process of parts production, assembly, testing, repair, delivery and sales is traced.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention identifies component defects before assembling finished products, controls the unqualified rate at the component level, and improves the qualified rate of finished products; 2. The present invention analyzes and repairs unqualified parts, improves product manufacturing process, and increases product qualification rate; 3. The present invention compiles unique codes for parts and components, and traces the production process of each part through the coding information. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of the steps of a method for automatically generating assembly traceability for a purely structural consumable material according to the present invention; Figure 2 A schematic diagram of a purely structural consumables automatic assembly traceability system according to the present invention; Figure 3 A flow chart of a method for automatically generating assembly traceability for pure structural consumables of the present invention for printing finished product codes and finished product packaging codes; Figure 4 Schematic diagrams of various codes for a method of automatically generating assembly traceability for a purely structural consumable material according to the present invention. DETAILED DESCRIPTION

[0018] 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 only 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.

[0019] Example: Figure 1-Figure 4 As shown, the present invention provides a technical solution, a method for automatically generating assembly traceability of pure structural consumables; After receiving the order, the matching table of finished product and component quality inspection results is captured from the finished product BOM, and the component code is assigned to the order according to the coding rules; The component code is to capture the matching table of finished products and component quality inspection from the finished product bill of materials, assign component codes to the order according to the customized coding rules, and generate an order matching table; the order matching table includes the finished product code, component code and quality inspection status; The generated order matching table is shown in the following table:

[0020] The table represents the matching table between the finished product PC-sn and the parts {SC1, SC2, ..., SCX}; The component code, each component has a unique component code; Use a laser engraving printer to print the part number in the designated area of ​​the part.

[0021] Conduct real-time quality inspection and marking of parts, and transmit the part code and quality inspection results back; Collect high-definition images of parts after printing and coding, and perform denoising, image enhancement and grayscale conversion on the collected high-definition images of parts; The denoising method uses a Gaussian filter to remove noise points in the image and reduce interference. The formula is as follows: ; Among them, G(x,y) represents the denoised image; w(i,j) represents the Gaussian weight; I(x,y) represents the original image; k represents the window radius of the filter; The image enhancement improves the quality of the image by contrast enhancement, and the formula is as follows: ; Among them, H(x,y) represents the enhanced image; μ represents the mean of the image; σ represents the standard deviation; The grayscale conversion uses a weighted average method to calculate the grayscale value of each pixel and convert the color image into a grayscale image; The processed images are input into the parts quality inspection model; The component quality inspection model includes a convolutional layer, a pooling layer and a fully connected layer; The convolution layer uses convolution kernels of different sizes to perform convolution operations on the image and extract local features. After the convolution operation is completed, the ReLU activation function is used to introduce nonlinearity so that the model can learn more complex features. The formula is as follows: F(X)=max(0,X); Among them, F(X) represents the output value of the activation function; X represents the input value of the activation function; The pooling layer, after the convolutional layer, is used to reduce the size of the feature map, reduce the amount of calculation, and retain key information; The fully connected layer, after the convolutional layer and the pooling layer, integrates the extracted features to form the output. The formula is as follows: Y = λ(W × X + b); Among them, Y represents the output value; λ represents the ReLU activation function; W represents the weight matrix; b represents the bias term; Use the Softmax activation function to convert the output into two categories: "qualified" and "unqualified". The formula is as follows: ; Among them, Y i Expressed as the predicted probability of category i; z i is represented as the original output of category i; C is represented as the total number of categories; j is represented as the index of the category.

[0022] During the training process, the component quality inspection model calculates the predicted value of each sample through forward propagation, compares it with the true label, and calculates the error through the loss function. The loss function formula is as follows: ; Among them, L represents the loss function; N represents the total number of samples; y i Represented as the true label of the i-th sample; represents the probability predicted by the component quality inspection model; The gradient of the loss relative to the network parameters is calculated by back propagation, and the optimizer is used to update the parameters of the parts quality inspection model. The formula is as follows: ; Among them, θ t Represents the parameters of the current component quality inspection model; η represents the learning rate, which controls the step size of parameter update; m t Expressed as the mean of the gradient; v tIt is represented as the variance of the gradient; ε is represented as a constant to prevent the denominator from being zero; The parameters of the component quality inspection model are updated once in each iteration until the component quality inspection model converges.

[0023] Through the component quality inspection model, it is determined whether the component is qualified; the components that pass the quality inspection are marked as "PASS", and the components that fail the quality inspection are marked as "NG", as shown in the following table:

[0024] Before the parts are ready for assembly, use image recognition equipment to read the unique coding information on the surface of each part and query the quality inspection results of the parts; when the quality inspection results of each part are qualified, that is, marked as "PASS", the assembly process will be entered; when the quality inspection results of any part are unqualified, that is, marked as "NG", the assembly process cannot be entered; After all parts and components are qualified in quality inspection and assembled into finished products, the assembled finished products will be assigned a unique finished product code, and the finished product code will be bound to the parts and components code.

[0025] The components that fail the quality inspection are analyzed and repaired; the repaired components are re-inspected for quality. If the quality inspection is qualified, the mark of the component is changed from "NG" to "PASS" and the process proceeds to step S4; if the component cannot be repaired, the component is classified as a defective product and cannot be assembled; For structural parts, if the quality inspection fails, it is necessary to inspect whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the failure of the quality inspection of the parts is due to appearance defects, the maintenance personnel will clean and repair the parts. After cleaning and repair, the quality inspection will be carried out again. If the quality inspection is qualified, the process proceeds to step S4, otherwise it is classified as a defective product and cannot be assembled; when the failure of the quality inspection of the parts is due to tolerances that do not meet the requirements, the tolerance data of the parts must be recorded and they will be classified as defective products and cannot be assembled.

[0026] The printing device obtains the finished product code and the packaging code, creates a timer, and sends the finished product code to the inkjet printer; after the inkjet printer receives the finished product code, when sensing the electrode signal, it starts printing the finished product code and sends a "start printing" signal to the printing device; after the inkjet printer completes printing, it returns the printed content to the printing device; After receiving the print content, the printing device turns off the timer and adds the packaging code to the print list; during the printing process of the inkjet printer, if the timer times out, a printing abnormality alarm is triggered; When the printing device receives the print completion signal from the cutter printer, the cutter printer will feed back the print content to the printing device; the printing device sends the packaging code to the cutter printer, and receives a successful reception confirmation message from the cutter printer; When the number of packaging codes in the print list reaches or exceeds m, where m is a positive integer set by the system; and after the sending and receiving operations of all print tasks have been completed, the printing device will remove the processed packaging code items from the queue.

[0027] After the finished product is packed, the packaging code and the box code are bound and delivered to the client; Through the coding information of parts and finished products, the entire production and quality inspection process can be traced back to the source, and the entire process of parts production, assembly, inspection and repair can be traced; through the coding information of parts, the delivery, sales and use process of the finished product can be reversely traced.

[0028] The present invention provides a technical solution, a pure structural consumables automatic generation assembly traceability system, the system includes an order receiving module, a component code printing module, a component quality inspection module, a finished product assembly and coding binding module, a packaging code and finished product code binding module, a non-conforming component analysis module and a traceability module; The order receiving module is used to receive the customer's order and extract the quality inspection matching table of finished products and parts from the finished product bill of materials; assign a part code to each order according to the customer's customized coding rules and generate an order matching table; the order matching table includes the finished product code, part code and quality inspection status; The component code printing module assigns a unique code to each component and sends the code of each component to the laser engraving printer, which prints the component code on a designated area on the surface of the component; The component quality inspection module collects high-definition images of components after printing and coding, performs denoising, image enhancement and grayscale conversion on the images, and then inputs them into the quality inspection model for judgment to determine whether the component quality inspection is qualified; components that pass the quality inspection are marked as "PASS" and components that fail the quality inspection are marked as "NG"; The finished product assembly and coding binding module uses an image recognition device to read the unique coding information on the surface of each component and query the component quality inspection results; after all components have passed the quality inspection, the finished product is assembled, the components that have passed the quality inspection are assembled, and a unique finished product code is assigned to the assembled finished product; after the assembly is completed, the finished product code is bound to the component code to ensure the transparency and traceability of the assembly process; the finished product assembly and coding binding module also includes finished product code and packaging code generation, finished product code printing function and packaging code printing function; The unqualified parts analysis module is used to conduct a detailed analysis of the parts that fail the quality inspection and determine the type of problem of the unqualified quality inspection; for repairable parts, they are repaired and re-inspected. If the quality inspection result is qualified, the mark of the part is changed from "NG" to "PASS", and the finished product assembly and coding binding module is re-entered; if it cannot be repaired, it is classified as a defective product and cannot enter the finished product assembly and coding binding module; for structural parts, if the quality inspection fails, it is necessary to check whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the unqualified quality inspection of the parts is due to appearance defects, the maintenance personnel will clean and repair the parts. After the cleaning and repair are completed, the quality inspection will be carried out again. If the quality inspection is qualified, the finished product assembly and coding binding module will be re-entered, otherwise it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; when the unqualified quality inspection of the parts is due to the tolerance not meeting the requirements, the tolerance data of the parts needs to be recorded, and it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; The traceability module records and stores the coding, production, assembly, testing, repair, delivery and sales information of parts and finished products; through the coding information of parts and finished products, the traceability of the entire production, quality inspection and sales process is completed, and the entire process of parts production, assembly, testing, repair, delivery and sales is traced.

[0029] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for automatically generating and assembling pure structural consumables and tracing the source, characterized by: The method comprises the steps of: S1. Receive orders and assign parts codes; S2. Print codes on parts; S3. Perform quality inspection and mark on parts and send back the code and quality inspection results; S4. Check the quality inspection results of parts and components, ensure that all parts and components are qualified before assembly, and bind the assembled finished product with the part code; S5. The finished product code is bound to the finished product packaging code; S6, analyzing and repairing the parts that failed the quality inspection. After the repaired parts pass the quality inspection, re-enter step S4; the parts that cannot be repaired are classified as unqualified products; S7. Complete traceability of finished products and parts through coded information.

2. A method for automatically generating and assembling pure structural consumables according to claim 1, characterized in that: The component code is to capture the matching table of finished products and component quality inspection from the finished product bill of materials, assign component codes to the order according to the customized coding rules, and generate an order matching table; the order matching table includes the finished product code, component code and quality inspection status; The component code, each component has a unique component code; Use a laser engraving printer to print the part number in the designated area of ​​the part.

3. A method for automatically generating and tracing the source of pure structural consumables according to claim 1, characterized in that: The quality inspection includes: Collect high-definition images of parts after printing and coding, and perform denoising, image enhancement and grayscale conversion on the collected high-definition images of parts; The denoising method uses a Gaussian filter to remove noise points in the image and reduce interference. The formula is as follows: ； Among them, G(x,y) represents the denoised image; w(i,j) represents the Gaussian weight; I(x,y) represents the original image; k represents the window radius of the filter; The image enhancement improves the quality of the image by contrast enhancement, and the formula is as follows: ； Among them, H(x,y) represents the enhanced image; μ represents the mean of the image; σ represents the standard deviation; The grayscale conversion uses a weighted average method to calculate the grayscale value of each pixel and convert the color image into a grayscale image; The processed image is input into the component quality inspection model, and the component quality inspection model is used to determine whether the component is qualified; components that pass the quality inspection are marked as "PASS", and components that fail the quality inspection are marked as "NG".

4. A method for automatically generating and tracing the source of purely structural consumables according to claim 3, characterized in that: The component quality inspection model includes a convolutional layer, a pooling layer and a fully connected layer; The convolution layer uses convolution kernels of different sizes to perform convolution operations on the image and extract local features. After the convolution operation is completed, the ReLU activation function is used to introduce nonlinearity so that the model can learn more complex features. The formula is as follows: F(X)=max(0,X); Among them, F(X) represents the output value of the activation function; X represents the input value of the activation function; The pooling layer, after the convolutional layer, is used to reduce the size of the feature map, reduce the amount of calculation, and retain key information; The fully connected layer, after the convolutional layer and the pooling layer, integrates the extracted features to form the output. The formula is as follows: Y = λ(W × X + b); Among them, Y represents the output value; λ represents the ReLU activation function; W represents the weight matrix; b represents the bias term; Use the Softmax activation function to convert the output into two categories: "qualified" and "unqualified". The formula is as follows: ； Among them, Y i Expressed as the predicted probability of category i; z i is represented as the original output of category i; C is represented as the total number of categories; j is represented as the index of the category.

5. A method for automatically generating and tracing the source of pure structural consumables according to claim 4, characterized in that: During the training process, the component quality inspection model calculates the predicted value of each sample through forward propagation, compares it with the true label, and calculates the error through the loss function. The loss function formula is as follows: ； Among them, L represents the loss function; N represents the total number of samples; y i Represented as the true label of the i-th sample; represents the probability predicted by the component quality inspection model; The gradient of the loss relative to the network parameters is calculated by back propagation, and the optimizer is used to update the parameters of the parts quality inspection model. The formula is as follows: ； Among them, θ t Represents the parameters of the current component quality inspection model; η represents the learning rate, which controls the step size of parameter update; m t Expressed as the mean of the gradient; v t It is represented as the variance of the gradient; ε is represented as a constant to prevent the denominator from being zero; The parameters of the component quality inspection model are updated once in each iteration until the component quality inspection model converges.

6. The method for automatically generating and tracing the source of pure structural consumables according to claim 1, characterized in that: Before the parts are ready for assembly, use image recognition equipment to read the unique coding information on the surface of each part and query the quality inspection results of the parts; when the quality inspection results of each part are qualified, that is, marked as "PASS", the assembly process will be entered; when the quality inspection results of any part are unqualified, that is, marked as "NG", the assembly process cannot be entered; After all parts and components are qualified in quality inspection and assembled into finished products, the assembled finished products will be assigned a unique finished product code, and the finished product code will be bound to the parts and components code.

7. The method for automatically generating and tracing the source of pure structural consumables according to claim 1, characterized in that: The components that fail the quality inspection are analyzed and repaired; the repaired components are re-inspected for quality. If the quality inspection is qualified, the mark of the component is changed from "NG" to "PASS" and the process proceeds to step S4; If a part cannot be repaired, it will be classified as defective and cannot be assembled; For structural parts, if the quality inspection fails, it is necessary to inspect whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the failure of the quality inspection of the parts is due to appearance defects, the maintenance personnel will clean and repair the parts. After cleaning and repair, the quality inspection will be carried out again. If the quality inspection is qualified, the process proceeds to step S4, otherwise it is classified as a defective product and cannot be assembled; when the failure of the quality inspection of the parts is due to tolerances that do not meet the requirements, the tolerance data of the parts must be recorded and they will be classified as defective products and cannot be assembled.

8. The method for automatically generating and assembling pure structural consumables according to claim 1, characterized in that: Before the finished product code is bound to the finished product packaging code, the finished product code and the finished product packaging code need to be printed; The printing device obtains the finished product code and the packaging code, creates a timer, and sends the finished product code to the inkjet printer; after the inkjet printer receives the finished product code, when sensing the electrode signal, it starts printing the finished product code and sends a "start printing" signal to the printing device; after the inkjet printer completes printing, it returns the printed content to the printing device; After receiving the print content, the printing device turns off the timer and adds the packaging code to the print list; During the printing process of the inkjet printer, if the timer times out, a printing abnormality alarm is triggered; When the printing device receives the print completion signal from the cutter printer, the cutter printer will feed back the print content to the printing device; The printing device sends the packaging code to the cutter printer and receives a successful reception confirmation message from the cutter printer; When the number of packaging codes in the print list reaches or exceeds m, where m is a positive integer set by the system; and after the sending and receiving operations of all print tasks have been completed, the printing device will remove the processed packaging code items from the queue.

9. The method for automatically generating and assembling pure structural consumables according to claim 1, characterized in that: After the finished product is packed, the packaging code and the box code are bound and delivered to the client; Through the coding information of parts and finished products, the entire production and quality inspection process can be traced back to the source, and the entire process of parts production, assembly, inspection and repair can be traced; through the coding information of parts, the delivery, sales and use process of the finished product can be reversely traced.

10. A purely structural consumables automatic generation and assembly traceability system, characterized by: The system includes an order receiving module, a component code printing module, a component quality inspection module, a finished product assembly and code binding module, a packaging code and finished product code binding module, a non-conforming component analysis module and a traceability module; The order receiving module is used to receive the customer's order and extract the quality inspection matching table of finished products and parts from the finished product bill of materials; Assign component codes to each order according to the customer's customized coding rules and generate an order matching table; the order matching table includes finished product codes, component codes and quality inspection status; The component code printing module assigns a unique code to each component and sends the code of each component to the laser engraving printer, which prints the component code on a designated area on the surface of the component; The component quality inspection module collects high-definition images of components after printing and coding, performs denoising, image enhancement and grayscale conversion on the images, and then inputs them into the quality inspection model for judgment to determine whether the component quality inspection is qualified; components that pass the quality inspection are marked as "PASS" and components that fail the quality inspection are marked as "NG"; The finished product assembly and coding binding module uses an image recognition device to read the unique coding information on the surface of each component and query the component quality inspection results; after all components have passed the quality inspection, the finished product is assembled, the components that have passed the quality inspection are assembled, and a unique finished product code is assigned to the assembled finished product; after the assembly is completed, the finished product code is bound to the component code to ensure the transparency and traceability of the assembly process; the finished product assembly and coding binding module also includes finished product code and packaging code generation, finished product code printing function and packaging code printing function; The unqualified parts analysis module is used to conduct a detailed analysis of the parts that fail the quality inspection and determine the type of problem of the unqualified quality inspection; repair the repairable parts and re-inspect them. If the quality inspection result is qualified, the mark of the parts is changed from "NG" to "PASS" and the finished product assembly and coding binding module is re-entered; If it cannot be repaired, it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; for structural parts, if the quality inspection fails, it is necessary to inspect whether there are defects in the appearance and whether the tolerance meets the specification requirements; when the quality inspection of parts fails due to appearance defects, the maintenance personnel will clean and repair the parts. After the cleaning and repair are completed, the quality inspection will be carried out again. If the quality inspection is qualified, it will re-enter the finished product assembly and coding binding module, otherwise it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; when the quality inspection of parts fails due to tolerance not meeting the requirements, the tolerance data of the parts needs to be recorded and it will be classified as a defective product and cannot enter the finished product assembly and coding binding module; The traceability module records and stores the coding, production, assembly, testing, repair, delivery and sales information of parts and finished products; through the coding information of parts and finished products, the traceability of the entire production, quality inspection and sales process is completed, and the entire process of parts production, assembly, testing, repair, delivery and sales is traced.

Images