Research On MGG Quality Inspection Of Automobile Seat Belt Based On Machine Vision

Strengthening the quality inspection of automobile parts and reducing the use of defective automobile parts and accessories are of great significance in solving traffic safety problems.Seat belts are important parts for protecting the safety of drivers and passengers during driving.The quality inspection of their spare parts are of great significance for ensuring the safety of seat belts and safety driving.At present,manual detection methods are mainly used to detect automobile parts.The detection cost is high and the efficiency is low.As an efficient non-contact detection technology,machine vision detection technology is superior to manual detection in terms of speed,accuracy,environmental adaptability,and cost.In this paper,Micro Gas Generator(MGG)quality inspection of Pre-tightened automobile seat belt in term surface defect and key location measurement was studied by machine vision.According to the detection requirement of MGG,the detection system was designed and built.This mainly includes: hardware and software analysis such as selection of imaging and lighting systems,analysis and selection of mechanical transmission part,analysis and selection of control software and hardware,design of control system wiring and the compilation of control program.For the detection of MGG surface defects,the defect detection and classification of normal,scratched and needle-less MGG was realized based on MATLAB design algorithm.Median filtering was used to denoised the MGG image;gray texture-gradient co-occurrence matrix was used to extract 15 texture feature parameters of MGG image;Principal Component Analysis(PCA)was used to reduce the feature size;A classifier based on RBF neural network was designed for classification and recognition.For the 105 samples tested,the false positive rate was1.94%.For the key dimension detection of MGG,the MATLAB design algorithm wasutilised to achieve the detection of its outer diameter and notch width.The image enhancement algorithm based on fuzzy set was also used to contrast the MGG image,and the binary image processing performed by OTSU to complete the image preprocessing.The edge detection of the MGG image was completed by Sobel differential operator;the Sobel operator eight-template linear interpolation subdivision algorithm detects the sub-pixel edge of the MGG and finds the coordinates of the edge points.Finally,the key dimension measurement of the MGG was performed using the least-squares method of straight line fitting and direct detection,and the measurement results show the detection accuracy meets the actual detection accuracy requirements.