Vehicle fault diagnosis with deep learning

At present, automobiles have become a common means of transportation, but with the increase of vehicles, safety issues have gradually emerged. Therefore, the assembly, manufacture and production of vehicles require systematic testing and rigorous inspection. Therefore, defect detection of vehicle parts is particularly important. Vehicle parts defect detection has evolved from manual detection of traditional classification methods to machine vision methods. In this paper, the deep learning method is used to firstly detect the defects of vehicle parts through the training of VGG۱۶ network structure model. The accuracy rate is ۹۴.۳۶. Secondly, the VGG۱۶ network structure model is improved. By introducing the inceptionv۳ module, the width of the model is increased on the basis of depth, the image is better recognized with an accuracy of ۹۵.۲۹. However, the accuracy of the traditional HOG+SVM classification method is only ۹۳.۸۸, and the efficiency of both methods is higher than that of the traditional method.