Research On Multi-uav Area Coverage Technology For Large Scale Panoramic Base Map Acquisition

With the increasing maturity of UAV technology,it is of great practical significance and wide application prospect to use UAV for area coverage aerial photography and to obtain panoramic base map of mission area.For a large area,it is time-consuming and inefficient to carry out area coverage mission with multiple sorties of a single UAV.Therefore,this thesis studies how to use multiple UAVs to jointly complete the area coverage mission and build a multi-UAV area coverage system to realize the practical application of the algorithm.The main research work of this thesis is as follows:Firstly,according to the structure characteristics and flight principle of the quadrotor UAV,a multi-UAV area coverage environment modeling algorithm based on the grid method is proposed,and the constraint conditions of multi-UAV area coverage mission assignment are established.On this basis,DARP algorithm is used for area coverage mission assignment of multiple UAVs,and it is improved to be suitable for area coverage mission assignment under the condition that the endurance performance parameters of multiple UAVs are different.The simulation results show that the mission assignment algorithm based on the improved DARP algorithm can obtain the mission assignment results that meet the mission assignment constraints when the number of UAVs,the initial position of UAVs,the shape and size of the mission area and the endurance performance parameters of UAVs are different,and the optimization time is short.Secondly,this thesis analyzes the shortcomings of two traditional coverage path planning algorithms(STC algorithm and wavefront algorithm)and improves them respectively.Then,an area coverage path planning algorithm based on DQN algorithm is proposed,and breadth first search algorithm and A* algorithm are used to help it escape from the dead zone.The simulation results show that the proposed algorithm based on DQN algorithm has the advantages of regular and orderly shape and less direction change compared with the two improved traditional algorithms.Thirdly,in order to obtain the panoramic base map of the mission area,this thesis studies the aerial image mosaic algorithm.On the basis of image feature-based method to realize aerial image registration,aiming at the complex and changeable overlapping area of aerial image with area coverage,a method of overlapping area division and mask establishment based on edge clipping is proposed.Based on this method,image pyramid fusion method is applied to aerial image fusion after registration to eliminate stitching gap and improve stitching effect.Finally,a multi-UAV area coverage system is built.The system can acquire the flight status data of multiple UAVs and the image data of pan tilt camera in real time,and control multiple UAVs.At the same time,the multi-UAV area coverage environment modeling,mission assignment,path planning and aerial image mosaic algorithm are encapsulated.The outfield experiment results show that the multi-UAV area coverage system built in this thesis can fully cover the mission area automatically by using multiple UAVs,and can stitch the aerial images of the area coverage to obtain the panoramic base map of the mission area.