Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Product appearance detection method based on cloud edge collaborative model optimization and implementation system thereof

A technology of appearance detection and modeling, which is applied in transmission systems, character and pattern recognition, instruments, etc., and can solve problems such as closed-loop optimization of data sets and models not involved, slow calculation speed, complex network mechanism, etc.

Pending Publication Date: 2021-05-11
SHANDONG UNIV
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The current industrial inspection system uses a relatively complex network mechanism, but the calculation speed is relatively slow, and it does not involve multi-target recognition and detection in a single picture, nor does it involve closed-loop optimization of data sets and models

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Product appearance detection method based on cloud edge collaborative model optimization and implementation system thereof
  • Product appearance detection method based on cloud edge collaborative model optimization and implementation system thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] A product appearance detection method based on cloud-edge collaborative model optimization. In this embodiment, the intelligent detection of an air conditioner external unit is used as an example to illustrate, as shown in figure 2 shown, including the following steps:

[0056] S1. Collect pictures of the appearance of air-conditioning external units that are known to be qualified, and after marking the pictures of the appearance of air-conditioning external units, establish a basic data set;

[0057] S2. Use the basic data set to train the YOLOv3-tiny model and the YOLOv3 model respectively, deploy the trained YOLOv3-tiny model on the edge server, and deploy the trained YOLOv3 model on the cloud platform; use the trained YOLOv3-tiny model as A Model, the trained YOLOv3 model is used as the B model.

[0058] S3. At the edge server, use the trained YOLOv3-tiny model to detect and recognize the picture of the appearance of the air conditioner external unit to be detecte...

Embodiment 2

[0070] A product appearance detection method based on cloud-edge collaborative model optimization according to Embodiment 1, the difference is that:

[0071] The detection method also includes step S6: using the basic data set obtained in step S5 to update, periodically update and train the YOLOv3-tiny model and the YOLOv3 model, and deploy the updated YOLOv3-tiny model to the edge device, after the update The YOLOv3 model is deployed on the cloud.

[0072] Periodically updating and training the YOLOv3-tiny model and the YOLOv3 model can improve the recognition accuracy, thereby realizing the self-optimization of the model. In the setting of the cycle, since the last update of the data set, the number of new pictures in the data set is n, and the update cycle is set to n 0 . Update the value of n when new pictures are added to the dataset, when n=n 0 The Shiyun platform starts to update and train the model, and resets n=0 at the same time, waiting for the next cycle.

[00...

Embodiment 3

[0086] A product appearance detection method based on cloud-edge collaborative model optimization according to Embodiment 1, the difference is that:

[0087] Socket network sockets are used for communication between the edge server and the cloud platform, the edge server is used as the client, and the cloud platform is used as the server for two-way communication.

[0088] In terms of communication between the edge device and the cloud platform, a stable and feasible socket network socket is used for communication. Socket communication first needs to create a socket between the client and the server to obtain the host name and port number; the server is used to monitor the request, after the client sends the communication request, the two establish a connection through a three-way handshake, and then send and receive after encoding , to complete the data transmission; after the transmission is completed, the client and the server wave four times to disconnect.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to a product appearance detection method based on cloud edge collaborative model optimization and an implementation system thereof. The detection method comprises the following steps: S1, establishing a basic data set; s2, training a YOLOv3-tiny model and a YOLOv3 model, and respectively deploying the YOLOv3-tiny model and the YOLOv3 model at an edge server and a cloud platform; s3, detecting and recognizing, by the YOLOv3-tiny model, the picture, when the picture is detected to be qualified, sending the picture and a detection resultto the cloud platform, and executing the S5, if the detection result is unqualified, determining that the result is suspected to be unqualified, sending the result to the cloud platform, and performing S4; s4, performing secondary detection on the suspected unqualified pictures on the cloud platform; if the secondary detection is not qualified, outputting a result, and ending the process, and if the secondary detection is qualified, performing S5; and S5, storing the qualified picture in the cloud platform, outputting a result, and ending. According to the method, the problems of accuracy, flexibility, time delay and data utilization rate are solved by using a working mode of cloud edge collaboration.

Description

technical field [0001] The invention relates to a product appearance detection method based on cloud-edge collaborative model optimization and its realization system, belonging to the technical field of edge computing architecture and artificial intelligence. Background technique [0002] The detection rate of industrial production lines is directly related to production efficiency. At present, most production lines use a fixed-model air-conditioning external unit identification method. The identification success rate is not high, and it is not easy to update and deploy the model. There are problems in the collection and processing of a large amount of industrial data. . [0003] The working mode of most existing factory production lines does not involve the cloud edge architecture, and there is still room for improvement in detection accuracy, problem analysis, and overall optimization. However, edge computing has high requirements for scene personalization, and its ration...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): H04L29/08G06K9/62
CPCH04L67/10H04L67/12G06F18/241G06F18/214
Inventor 张海霞马睿袁东风王翊州
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com