Surface defect detection for wire ropes based on deep convolutional neural network

Real-time, efficient detection of ire ropes (WR) surface detect is a challenging task. Affected by the performance of the algorithm, there is a problem that the detection and diagnosis effects are not ideal. To this end, an automatic detection method for surface detects of WRs on the basis of deep convolutional neural networks (DCNN) is put forward in this article. First, according to the actual situation, the state of the WR is defined as three main states of health, broken wire and wear. Then, a big data set of the WR surface image with three states is established, and an improved LAWN model is used to perform deep mining on the established data set. Finally, performance comparisons are made with traditional machine learning algorithms. Through a large number of tests and contrasting analysis, the results show that the method we proposed can achieve a higher diagnostic accuracy than the traditional methods, which meets the actual detection requirements.