Quality-Controllable Viewpoint Planning For Efficient Aerial Inspection Of Wind Power Equipment

Unmanned aerial vehicle(UAV)inspection of wind power equipment surfaces is an important part of operations and maintenance,which is mainly done by UAVs carrying sensors to fly near the wind power equipment and collect images to inspect for any defects.Therefore,effective planning of the UAV shooting behavior is of great significance to wind power inspection.In autonomous inspection by drones,in order to obtain high-quality imaging data that covers large and complex objects to be inspected,a large number of waypoints are generally designed,which greatly affects the operational flight time and limits the inspection efficiency,seriously restricting the large-scale industrial application of drones.To address these issues,this paper proposes a new viewpoint planning algorithm based on continuous differentiable sampling for known structural inspection equipment to reduce the number of waypoints,i.e.,a multi-directional viewpoint planning algorithm for UAVs based on continuous differentiable sampling.The algorithm achieves the goal of significantly reducing the number of waypoints while satisfying the requirements of comprehensive coverage of inspection targets and data acquisition quality.This paper conducts scientific research in multiple aspects such as optimized viewpoint sampling,optimized waypoint selection,and zoom planning for multiple viewpoints.Firstly,the target model for mathematical modeling of viewpoint optimization is determined,the description form of the viewpoint is defined,and the gimbal pitch angle,yaw angle,and camera imaging geometry are used as constraints to establish a continuous differentiable view quality objective function for viewpoint collection,and to construct a mathematical model for viewpoint sampling optimization.Secondly,for the first time,a method for drone planning multiple viewpoint directions in a single waypoint is proposed,and an objective function for minimizing the number of waypoints is established.Combined with the inspection coverage rate requirements,the number of waypoints is optimized,and a greedy algorithm is used to optimize a small number of high-quality candidate viewpoint spatial positions,to obtain the waypoint positions under the condition of full coverage of the inspection targets.Finally,zoom planning is carried out for multiple viewpoint directions of each waypoint to meet the accuracy requirements of different inspection technical specifications.To demonstrate the effectiveness of the proposed method,this paper compares the number of waypoints obtained by planning for different inspection models and the rate of reduction in the number of waypoints.Simulation data of the collected images are also presented.Compared with other excellent methods in the field,this paper demonstrates that the proposed method can reduce the number of waypoints by at least 77%,greatly improving the data acquisition efficiency of drone operations and achieving quality controllable viewpoint planning.