UAV patrol path planning based on machine vision and multi-sensor fusion

Abstract With the rapid development of unmanned aerial vehicle (UAV) technology, there are more and fields UAV application. This research mainly discusses patrol path planning based on machine vision multi-sensor fusion. article studies how to apply ultrasonic, a classic ranging sensor, obstacle avoidance UAVs. The designed ultrasonic system is complete set hardware software systems. part consists forward module central signal processing module. Among them, single-axis stabilization gimbal for module, which decouples attitude angle pitch detection sensor. In Kalman filtering performed data in four directions front, rear, left, right, control sent flight controller according filtered sensor data. At same time, human–computer interaction interface also various parameters system. For route method tower, routine steps used inspect tower with single-circuit line, specific targets insulator string, ground wire, conductor. this study, average statistical result straight-line distance patrolling 100 m 99.80 m, error only 0.2%. fusion suitable engineering application perception avoidance. adopted can be extended most platforms, it broad prospects.