Research on brake pad surface defects detection based on deep learning

In order to solve the problem of feature extraction of brake pad surface defects in traditional machine learning, an automatic defect detection system based on deep learning is designed. Firstly, the image acquisition system is designed and the brake pad image dataset is constructed. Then, a kind of defect detection method based on convolutional neural network (CNN) is proposed. The dataset is input into the CNN model for training, and the appropriate defect classification model of brake pad is selected for defect detection and classification. In addition, in order to further improve the performance of the detection system, a defect detection method based on full convolution network (FCN) is also proposed. Finally, the two methods mentioned above are tested. The results show that both methods are suitable for the detection of brake pad defects and can realize the recognition and classification of brake pad defects. In comparison, the method based on CNN has higher detection efficiency than the one based on FCN, but the method based on FCN has better recognition accuracy and stronger performance than the one based on CCN.