Research On Key Technology And Testing Device Of Visual Inspection For Synthetic Resin Glove

Visual detection technology has developed rapidly in recent years.Its core is to use machine vision instead of human eyes to measure and judge the target.It has great application value in the production and manufacturing industry to prevent the inflow of defective products into the market.This research is aimed at detecting the defect of disposable synthetic resin gloves,which is popular in the market,and studies a visual based defect detection method and testing device for synthetic resin gloves.This paper focuses on the detection methods of plastic gloves,fingers,fingers and surface contamination,and the key technologies of detection devices.First of all,according to the defect detection method,the synthetic resin gloves adactylism finger detection program and detection process.Based on the application of visual detection technology in other products detection and the basic requirements of detecting the defects of gloves in this paper,a multi threshold improved Canny algorithm for edge detection of synthetic resin gloves was proposed.Among them,multiple thresholds of the histogram fitting curve are the highlight of the method.By replacing the artificial threshold of Canny algorithm with multiple thresholds,the accuracy of edge detection is improved.By using various algorithms to detect the edge of synthetic resin gloves,the feasibility and superiority of the algorithm are verified by comparing the detection results.Second,in view of the classification and identification of oil pollution defects on the surface of synthetic resin gloves,the detection methods of surface defects in many kinds of products are studied,and the detection methods of oil pollution defects on the surface of gloves with HOG features and SVM are determined.This paper mainly analyzes the basic theory of image HOG feature extraction,and analyzes the flow and theoretical basis of using SVM to create classifier.By collecting the positive and negative sample images of the glove,the MATLAB is used to verify the method,and the feasibility and accuracy of the classification recognition are analyzed,and the better classification results are obtained.Third,referring to the working mode of the existing synthetic resin gloves production line,we have developed the synthetic resin gloves detection device and the design detection device control system.The device mainly simulates the movement process of the synthetic resin gloves in the production,provides the reasonable position of the glove in the image collection,gets the positive and anti two faces of the glove,designs the components of the control system module and the control logic,and designs the human-computer interaction interface for the convenient operation.Finally,the hardware and software system of the synthetic resin glove defect detection is established,and the experimental platform is built to carry out the experiment.The key technologies of the hardware and software systems are studied.The principles and theoretical models of the LED light source system and the parameters,precision,working flow and image acquisition advantages of the Bumblebee2 binocular camera are emphatically analyzed.The MATLAB detection interface with two defects of edge and oil pollution is created to achieve real-time image pre-processing and real-time display in the detection process.