Research on Automatic Target Recognition of UAV Power Inspection Image Based on Deep Learning Intelligent Algorithm

With the development of sensor technology, network technology and communication technology, autonomous unmanned system has made great progress in recent years. At present, drones have become an indispensable part of human life: Electricity Inspection, map survey, air traffic control, film production, logistics and transportation, marine environmental monitoring, marine surveillance and search and rescue. Every application of UAV system brings great convenience to our life and improves the convenience of our life. In recent years, more and more uavs provide effective support for electric power safety patrol inspection, which has become the current research hotspot. In this paper, based on depth learning algorithm, using UAV image data, considering UAV image characteristics and UAV resource constraints, a target detection method suitable for UAV applications is implemented, aiming at the characteristics of small ratio of target pixels in UAV image, the structure of algorithm network is improved, and the detection algorithm for small and weak characteristic target in UAV image is realized, the algorithm is suitable for UAV image recognition during power inspection.