Design Of Socks Detection And Sorting System Based On Machine Vision

Machine vision inspection is a kind of inspection technology that integrates machinery,automation,computer image processing and other technologies.The technology is widely used in all walks of life,mainly manifested in product quality identification and product surface defect detection.This paper uses machine vision detection and sorting technology to realize the intelligent detection and sorting task of the hosiery process,which not only has theoretical research significance but also improves the production efficiency and capacity of traditional hosiery enterprises.This paper is based on the machine vision detection and sorting technology,and takes the socks produced by the actual socks enterprises as the detection object,and designs a set of intelligent machine vision technology based sock detection and sorting system.The main research contents and results are as follows:First,analyze the functions required by the entire system,and design the architecture of the entire system based on the required functions.Each subsystem is designed by the design scheme of the entire system,such as: vision system,sorting system,transmission system,image processing system,control system.Then complete the system hardware platform construction.Secondly,to study the algorithm of sock image recognition.The algorithm adopted in this paper is the deep learning neural network algorithm,and the deep learning algorithm is used to process the sock samples,train the model,and predict whether the sock is qualified.The model is trained by this algorithm to predict socks,so that the detection accuracy of the entire system reaches about 87%,which meets the requirements of enterprises.Therefore,the research content is also the core work of this article.Then,the software system part of the detection system was perfected.The software system is mainly divided into login and registration system and hardware information query and sock image detection system.The login and registration system mainly protects the security of the system.The hardware information query and sock image detection system can mainly check the hardware information of the entire system and the time of operation of the system to detect the sock picture qualification.Finally,the sock image detection part and sorting part of the entire machine vision sock detection and sorting system are tested and tested,and the detection accuracy and sorting accuracy of the system are obtained based on the test data and the evaluation criteria.