Abstract:The failure of “three-cross” transmission lines will have a major impact on the basic transportation facilities such as power security and railways and highways. Therefore, it is crucial to identify and warn of the potential risks of “three-cross” transmission lines. In response to this problem, this paper proposes a method combined with deep learning and image processing technology. It can be used to identify and detect engineering machinery equipments in the safe area of transmission lines. The algorithm uses the regional convolutional neural network (Faster-RCNN) to identify and detect various types of engineering machinery equipment located in the safe area of the transmission line, and implements it based on the caff framework. The algorithm also combines image processing techniques such as homomorphic filtering to further improve the accuracy of target detection results in complex environments such as poor illumination. The results of multiple experiments show that the algorithm has a high detection and recognition rate for all kinds of engineering machinery and equipment in various complex environments. The algorithm realizes the early warning and identification of the risk of engineering machinery located in the safe area of the transmission line. It provides a basis for the establishment of the intelligent identification platform of the transmission line fault and risk.