基于深度学习的隧道病害识别方法研究

iseases in railway tunnels always threaten the safety of railway operation. The automatic and accurate recognition of diseases by using computer vision technology has become a hot research topic. Common tunnel diseases include cracks, water leakage, block dropping, scratches and so on. There are many difficulties in recognizing these tunnel diseases, such as the diseases are always too small to berecognized, a variety of diseases appear at the same time, diseasesare not obvious, which brings great difficulty in actual recognizing. In this paper, we propose a deep learning method to automaticallyrecognize of tunnel diseases, and we build a tunnel disease dataset in order to train the model.Experiments have proved that our method achieves a good performance in tunnel disease recognizing.