Multi-directional viewpoints planning of UAV inspection based on continuous differentiable sampling
When Unmanned Aerial Vehicle(UAV)is applied to the inspection of electric equipment,it is very impor-tant to improve the inspection efficiency and the quality of inspection data.Thus,a new Multi-Directional View Point Planning algorithm(MD-VPP)for the inspection of known complex equipment structures was present,which could greatly reduce the number of waypoints while satisfying the requirement of comprehensive coverage inspection target and data acquisition quality.A view quality function of continuous and differentiable view acquisition was con-structed with the pitch,yaw and camera projection of PTZ as constraints,and the candidate view sets was opti-mized,and then the waypoint pose of the coveraged object was solved by greedy algorithm.The number and reduc-tion rate of waypoints under different patrol objects were compared and analyzed experimentally.Compared to other excellent methods in the field,the proposed method reduced the number of waypoints by at least 77％while ensuring full coverage and greatly improved the data acquisition efficiency of a single UAV operation.
NETL
NSTL
万方数据
