A method and system for comprehensive management of industrial information based on Internet of Things technology

By adopting IoT technology and defect detection models on industrial production lines, automated inspection and packaging products are solved, and automated production and inspection are realized, management efficiency is improved and costs are reduced.

CN118691055BActive Publication Date: 2024-12-03C-TOP ELECTRONICS CO LTD
View PDF 12 Cites 0 Cited by

Patent Information

Application Number
CN202411183825.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-12-03
Estimated Expiration
2044-08-27

AI Technical Summary

Technical Problem

Industrial products require manual information statistics and quality inspection during production management, resulting in low efficiency and high cost, and error-prone. The existing technology has failed to effectively solve these problems.

Method used

Using the comprehensive industrial information management method based on the Internet of Things technology, the batch information of raw materials on the production line is collected, and the product is tested in the production process, and the detection records are generated. The mutual information method and the Pearson coefficient method are used to construct defect detection. Model, perform automated inspection and packaging decisions.

Benefits of technology

It realizes the automated production, inspection and packaging of industrial products, improves management efficiency, reduces labor costs, and reduces the factory risk of product defects through automated inspection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118691055B_ABST
    Figure CN118691055B_ABST
Patent Text Reader

Abstract

The present invention discloses an industrial information comprehensive management method and system based on the Internet of Things technology, and relates to the field of Internet of Things technology. The method includes collecting batch information of raw materials used on industrial production lines; performing station detection on products on industrial production lines in the production process link, and generating station detection records; obtaining the station detection records of products before the product production is completed and enters the factory packaging link; judging whether the station detection records of products meet preset conditions; and when the station detection records of products meet the preset conditions, the products are packaged for delivery. The present application realizes the automated production, automated detection and automated packaging of industrial products on industrial production lines, improves the management efficiency of industrial production lines, and reduces the labor cost of managing industrial production lines.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of Internet of Things technology, and more specifically, to an industrial information comprehensive management method and system based on Internet of Things technology. Background Art

[0002] Industrial products require manual information statistics screening and quality inspection during the production management process, but manual operation is not only prone to errors, but also causes damage to parts. It is understandable that manual management and operation is not only inefficient, but also too costly.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide an industrial information comprehensive management method and system based on Internet of Things technology to solve the above-mentioned technical problems.

[0005] The present application provides an industrial information comprehensive management method based on Internet of Things technology, which collects batch information of raw materials used on industrial production lines; performs station-passing detection on products on the industrial production line in the production process link, and generates station-passing detection records; obtains the station-passing detection records of the products before the production of the products is completed and enters the factory packaging link; determines whether the station-passing detection records of the products meet preset conditions; and performs factory packaging on the products when the station-passing detection records of the products meet the preset conditions.

[0006] Furthermore, the passing station detection of the products on the industrial production line in the production process link and the generation of the passing station detection record include: collecting the first panoramic video data stream of the products on the industrial production line in the production process link at the basic production process detection site and the second panoramic video data stream at the current production process detection site through a shooting device; performing distortion correction on the first panoramic video data stream and the second panoramic video data stream to obtain a first corrected video stream and a second corrected video stream; marking the key defect points of the products in the first corrected video stream and the second corrected video stream to construct a multi-classification training data set; training a defect detection model through the multi-classification training data set, wherein the defect detection model is constructed based on the mutual information method and the Pearson coefficient method; and inputting the first corrected video stream into the trained defect detection model to generate the passing station detection record.

[0007] Furthermore, the defect detection model is trained by the multi-classification training data set, including: decomposing the multi-classification training data set into n frames of static images corresponding to n video stream information; wherein n is a positive integer; extracting time series features from the n frames of static images based on the mutual information method and the Pearson coefficient method to obtain a time series feature vector; predicting the time series feature vector based on a two-stage time series attention mechanism to obtain a time series prediction sequence; optimizing the loss function of the defect detection model with the time series prediction sequence; the parameters of the defect detection model include a weight matrix and a bias, wherein the optimization training process of the defect detection model is: setting the current optimization training times of the defect detection model to d, and the maximum training times to , where the initial value of d is 0; set the loss function Loss for the d+1th optimization training:

[0008]

[0009] Wherein, m is the number of sequences in the time series prediction sequence, Indicates the expected output, represents the output value after training, n represents the total number of connection weights, represents the weight matrix of the feedforward neural network, Represents the error rate of the defect detection model after the d-th optimization training; the defect detection model obtained by training is optimized based on a recurrent neural network.

[0010] Furthermore, the defect detection model obtained through training is optimized based on the recurrent neural network, including: establishing a backbone network of the recurrent neural network; establishing a multi-scale fusion module of the recurrent neural network; establishing a classification head module of the recurrent neural network; inputting the n frames of static images into the classification head module for feature extraction to obtain n feature vectors; performing differential sampling on the n feature vectors using a differential sampling strategy; optimizing and training the defect detection model with the n feature vectors obtained through the differential sampling; wherein the differential sampling of the n feature vectors using the differential sampling strategy includes: calculating the imbalance rate corresponding to the n feature vectors , where K represents the number of defect key points contained in the n feature vectors, Indicates the number of defect key points not included in the n feature vectors; respectively calculate the distance between any two vectors in the n feature vectors, and calculate the local density of the defect feature vectors in the n feature vectors based on the following expression:

[0011]

[0012] in, Defect feature vector The local density of; dis represents the distance threshold; Indicates the n feature vectors and the defect feature vector The number of vectors whose distance is within the range dis.

[0013] Furthermore, the learning rate is set to 0.02, the training is optimized to obtain the optimal model parameters, and the trained defect detection model is constructed based on the optimal model parameters.

[0014] Furthermore, when the transit detection record of the product does not satisfy the preset condition, the product is intercepted.

[0015] Furthermore, after the product is packaged for delivery, the method further includes: reporting the information of the product to an industrial identification resolution node platform to implement industrial identification resolution of the product.

[0016] The present application also provides an industrial information integrated management system based on the Internet of Things technology, including a raw material information collection unit, a station inspection record generation unit, a station inspection record acquisition unit, and a factory packaging unit; wherein,

[0017] The raw material information collection unit is used to collect batch information of raw materials used on the industrial production line; the station-passing detection record generation unit is used to perform station-passing detection on products on the industrial production line in the production process link and generate station-passing detection records; the station-passing detection record acquisition unit is used to obtain the station-passing detection record of the product before the product is produced and enters the factory packaging link; the factory packaging unit is used to determine whether the station-passing detection record of the product meets the preset conditions; if the station-passing detection record of the product meets the preset conditions, the product is factory packaged.

[0018] The present invention further provides an electronic device, comprising:

[0019] a memory storing at least one instruction; and

[0020] The processor executes the instructions stored in the memory to implement the above-mentioned industrial information comprehensive management method based on the Internet of Things technology.

[0021] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned industrial information comprehensive management method based on Internet of Things technology.

[0022] Based on the embodiments provided by this application, the batch information of raw materials used on the industrial production line is collected; the products on the industrial production line in the production process are inspected and the inspection records are generated; before the product is produced and enters the factory packaging link, the inspection records of the products are obtained; it is determined whether the inspection records of the products meet the preset conditions; if the inspection records of the products meet the preset conditions, the products are packaged for delivery. Thus, the automated production, automated inspection and automated packaging of industrial products on the industrial production line are realized, the management efficiency of the industrial production line is improved, and the labor cost of managing the industrial production line is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the embodiments of the present invention and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 It is a flowchart of an optional industrial information integrated management method based on Internet of Things technology according to an embodiment of the present application;

[0025] Figure 2 It is a flowchart of another optional industrial information integrated management method based on Internet of Things technology according to an embodiment of the present application;

[0026] Figure 3 This is a structural diagram of an optional industrial information integrated management system based on Internet of Things technology according to an embodiment of the present application.

[0027] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0029] like Figure 1 As shown, the present application provides an industrial information integrated management method based on Internet of Things technology, including:

[0030] S101, collecting batch information of raw materials used in industrial production lines;

[0031] S102, performing station inspection on products on the industrial production line in the production process link and generating station inspection records;

[0032] S103, before the product is produced and enters the factory packaging stage, obtain the product's passing station inspection record;

[0033] S104, determining whether the product's station inspection record meets a preset condition;

[0034] S105, when the product's station inspection record meets the preset conditions, the product is packaged for delivery.

[0035] Based on the embodiments provided by this application, the batch information of raw materials used on the industrial production line is collected; the products on the industrial production line in the production process are inspected and the inspection records are generated; before the product is produced and enters the factory packaging link, the inspection records of the products are obtained; it is determined whether the inspection records of the products meet the preset conditions; if the inspection records of the products meet the preset conditions, the products are packaged for delivery. Thus, the automated production, automated inspection and automated packaging of industrial products on the industrial production line are realized, the management efficiency of the industrial production line is improved, and the labor cost of managing the industrial production line is reduced.

[0036] Furthermore, if Figure 2 As shown, the products on the industrial production line in the production process are inspected and the inspection records are generated, including:

[0037] S201, collecting, by means of a camera, a first panoramic video data stream of a product on an industrial production line in a production process at a basic production process detection site and a second panoramic video data stream at a current production process detection site;

[0038] S202, performing distortion correction on the first panoramic video data stream and the second panoramic video data stream to obtain a first corrected video stream and a second corrected video stream;

[0039] S203, marking the key points of defects of the product in the first corrected video stream and the second corrected video stream to construct a multi-classification training data set;

[0040] S204, training a defect detection model using a multi-classification training data set, wherein the defect detection model is constructed based on a mutual information method and a Pearson coefficient method;

[0041] S205: Input the first corrected video stream into the trained defect detection model to generate a transit detection record.

[0042] Furthermore, the defect detection model is trained through a multi-classification training data set, including:

[0043] Decompose the multi-classification training data set into n frames of static images corresponding to n video stream information; where n is a positive integer;

[0044] Based on the mutual information method and the Pearson coefficient method, the time series features of n frames of static images are extracted to obtain the time series feature vector;

[0045] The time series feature vector is predicted based on the two-stage time series attention mechanism to obtain the time series prediction sequence;

[0046] The loss function of the defect detection model is optimized with the time series prediction sequence; the parameters of the defect detection model include the weight matrix and the bias, wherein the optimization training process of the defect detection model is:

[0047] Set the current optimization training times of the defect detection model to The maximum number of training times is , where the initial value of d is 0; set the loss function Loss for the d+1th optimization training:

[0048]

[0049] Among them, m is the number of sequences in the time series prediction sequence, Indicates the expected output, represents the output value after training, n represents the total number of connection weights, represents the weight matrix of the feedforward neural network, represents the error rate of the defect detection model after the d-th optimization training;

[0050] The trained defect detection model is optimized based on recurrent neural network.

[0051] Furthermore, the trained defect detection model is optimized based on the recurrent neural network, including:

[0052] Establish a backbone network of a recurrent neural network; establish a multi-scale fusion module of a recurrent neural network; establish a classification head module of a recurrent neural network; input n frames of static images into the classification head module for feature extraction respectively, and obtain n feature vectors;

[0053] The n feature vectors are differentially sampled using a differential sampling strategy; the defect detection model is optimized and trained using the n feature vectors that have undergone differential sampling; wherein the n feature vectors are differentially sampled using a differential sampling strategy including:

[0054] Calculate the imbalance rate corresponding to n eigenvectors , where K represents the number of defect key points contained in n feature vectors, Indicates the number of defect key points that are not included in the n feature vectors;

[0055] The distance between any two vectors in the n eigenvectors is calculated respectively, and the local density of the defect eigenvectors in the n eigenvectors is calculated based on the following expression:

[0056]

[0057] in, Defect feature vector The local density of; dis represents the distance threshold; Represents the defect feature vector among n feature vectors The number of vectors whose distance is within the range dis.

[0058] Furthermore, the learning rate is set to 0.02, the optimal model parameters are obtained through optimization training, and the trained defect detection model is constructed based on the optimal model parameters.

[0059] Furthermore, when the product's transit inspection record does not meet preset conditions, the product is intercepted.

[0060] Furthermore, after the product is packaged for delivery, the method further comprises:

[0061] Report the product information to the industrial identification resolution node platform to realize the industrial identification resolution of the product.

[0062] like Figure 3 As shown, the present application provides an industrial information integrated management system based on the Internet of Things technology, which implements the above-mentioned industrial information integrated management method based on the Internet of Things technology. The system includes a raw material information collection unit 301, a station-passing detection record generation unit 302, a station-passing detection record acquisition unit 303 and a factory packaging unit 304; wherein,

[0063] The raw material information collection unit 301 is used to collect batch information of raw materials used in industrial production lines;

[0064] The station-passing detection record generating unit 302 is used to perform station-passing detection on products on the industrial production line in the production process link and generate station-passing detection records;

[0065] The station inspection record acquisition unit 303 is used to acquire the station inspection record of the product before the product is produced and enters the factory packaging stage;

[0066] The factory packaging unit 304 is used to determine whether the product's station inspection record meets the preset conditions; if the product's station inspection record meets the preset conditions, the product is factory packaged.

[0067] Based on the embodiments provided by this application, the batch information of raw materials used on the industrial production line is collected; the products on the industrial production line in the production process are inspected and the inspection records are generated; before the product is produced and enters the factory packaging link, the inspection records of the products are obtained; it is determined whether the inspection records of the products meet the preset conditions; if the inspection records of the products meet the preset conditions, the products are packaged for delivery. Thus, the automated production, automated inspection and automated packaging of industrial products on the industrial production line are realized, the management efficiency of the industrial production line is improved, and the labor cost of managing the industrial production line is reduced.

[0068] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A comprehensive industrial information management method based on Internet of Things technology, characterized in that: include: Collect batch information of raw materials used in industrial production lines; Conducting station inspection on products on the industrial production line in the production process and generating station inspection records; Before the product is produced and enters the factory packaging stage, obtain the product's transit inspection record; Determine whether the station-passing inspection record of the product meets preset conditions; If the station-passing inspection record of the product meets the preset condition, the product is packaged for delivery; The step of performing station inspection on the products on the industrial production line in the production process and generating station inspection records includes: Collecting, by means of a camera, a first panoramic video data stream of products on the industrial production line in a production process at a basic production process detection site and a second panoramic video data stream at a current production process detection site; Performing distortion correction on the first panoramic video data stream and the second panoramic video data stream to obtain a first corrected video stream and a second corrected video stream; Annotate the key points of defects of the product in the first corrected video stream and the second corrected video stream to construct a multi-classification training data set; Training a defect detection model using the multi-classification training data set, wherein the defect detection model is constructed based on a mutual information method and a Pearson coefficient method; Inputting the first corrected video stream into the trained defect detection model to generate the station-passing detection record; The step of training the defect detection model using the multi-classification training data set includes: Decomposing the multi-classification training data set into n frames of static images corresponding to n video stream information; wherein n is a positive integer; Extracting time series features from the n frames of static images based on the mutual information method and the Pearson coefficient method to obtain a time series feature vector; Predicting the time series feature vector based on a two-stage time series attention mechanism to obtain a time series prediction sequence; The loss function of the defect detection model is optimized with the time series prediction sequence; the parameters of the defect detection model include a weight matrix and a bias, wherein the optimization training process of the defect detection model is: Set the current optimization training times of the defect detection model to The maximum number of training times is , where the initial value of d is 0; Set the loss function Loss for the d+1th optimization training: ; Wherein, m is the number of sequences in the time series prediction sequence, represents the expected output of the ith sequence, represents the output value of the i-th sequence after training, n represents the total number of connection weights, represents the weight matrix of the jth feedforward neural network, represents the error rate of the defect detection model after the d-th optimization training; Optimizing the defect detection model obtained through training based on a recurrent neural network; The optimizing the defect detection model obtained by training based on the recurrent neural network includes: Establishing the backbone network of the recurrent neural network; establishing the multi-scale fusion module of the recurrent neural network; establishing the classification head module of the recurrent neural network; inputting the n frames of static images into the classification head module for feature extraction respectively, and obtaining n feature vectors; The n feature vectors are differentially sampled using a differential sampling strategy; the defect detection model is optimized and trained using the n feature vectors that have undergone the differential sampling; wherein the n feature vectors are differentially sampled using a differential sampling strategy, including: Calculate the imbalance rate corresponding to the n eigenvectors , where K represents the number of defect key points contained in the n feature vectors, Indicates the number of defect key points not included in the n feature vectors; The distance between any two vectors in the n feature vectors is calculated respectively, and the local density of the defect feature vectors in the n feature vectors is calculated based on the following expression: ; in, Defect feature vector The local density of; dis represents the distance threshold; Indicates the n feature vectors and the defect feature vector The number of vectors whose distance is within the range dis.

2. The method for comprehensive industrial information management based on Internet of Things technology according to claim 1 is characterized in that: The learning rate is set to 0.02, the optimal model parameters are obtained through optimization training, and the defect detection model after training is constructed based on the optimal model parameters.

3. The method for comprehensive industrial information management based on Internet of Things technology according to claim 1 is characterized in that: When the station-passing detection record of the product does not satisfy the preset condition, the product is intercepted.

4. The method for comprehensive industrial information management based on Internet of Things technology according to claim 1 is characterized in that: After the product is packaged for delivery, the method further comprises: The information of the product is reported to the industrial identification resolution node platform to implement industrial identification resolution of the product.

5. An industrial information integrated management system based on Internet of Things technology, the system implements the method as described in any one of claims 1 to 4, characterized in that: It includes a raw material information collection unit, a station inspection record generation unit, a station inspection record acquisition unit and a factory packaging unit; among which, The raw material information collection unit is used to collect batch information of raw materials used in industrial production lines; The station-passing detection record generating unit is used to perform station-passing detection on the products on the industrial production line in the production process link and generate a station-passing detection record; The station-passing inspection record acquisition unit is used to acquire the station-passing inspection record of the product before the product is produced and enters the factory packaging stage; The factory packaging unit is used to determine whether the station-passing detection record of the product meets a preset condition; if the station-passing detection record of the product meets the preset condition, the product is factory packaged.

Citation Information

Patent Citations

  • Article recommendation method based on generalized nerve attention

    CN111127142A

  • Financial time series prediction method based on two-stage attention mechanism generative adversarial network

    CN111475546A

  • Image detection and analysis method based on visual computing video transmission

    CN111860327A

  • Industrial product defect detection method and device

    CN115272340A

  • Monitoring method for dealing with webpage faults and webpage vulnerabilities

    CN115314267A