Gear grinding burn segmentation method based on deep convolutional neural network

The invention discloses a gear grinding burn segmentation method based on a deep convolutional neural network, and the method comprises the steps: obtaining a gear tooth side surface image of a to-be-detected gear, and carrying out the data preprocessing; amplifying the preprocessed data by using a data enhancement technology, and making a gear tooth side surface grinding burn segmentation data set; a gear grinding burn segmentation model based on the improved U-Net network is constructed and trained; inputting to-be-tested data, extracting and visualizing a gear tooth side surface grinding burn area, and judging the classification and level of gear tooth side surface grinding burn according to the color depth and area proportion of the grinding burn area. According to the method, an end-to-end gear grinding burn segmentation model is constructed, shallow high-resolution local features extracted by a convolutional neural network and sequence global features extracted by a self-attention mechanism are fused, t