Research On UAV Path Planning And Related Algorithms Based On Stereo Vision

With the continuous development of UAV and related technologies,UAV has been widely used in military and civilian fields.In order to make the UAV perform tasks more intelligently,it is necessary to solve the positioning and navigation problems of the UAV itself.In the early days,UAV mainly used lidar,GPS,IMU and so on for positioning.With the development of machine vision technology,stereo vision technology is widely used in the field of autonomous navigation of UAV.This paper focuses on the stereo vision SLAM process,UAV flight control algorithm and path planning algorithm,and completed the corresponding simulation experiments.Firstly,this paper designs a three-tier structure with service application layer,UAV decision layer and hardware execution layer,and a set of visual SLAM scheme for UAV navigation is designed on the basis of ORB-SLAM framework.In the visual front end,the image moment is used to improve the ORB feature extraction algorithm,which improves the efficiency of ORB feature extraction.In order to solve the problem that the traditional framework can't reconstruct dense 3D map,a method of real-time 3D reconstruction in the form of octree map using key frame information is proposed.Secondly,in the aspect of UAV path planning,in order to improve the efficiency of path search,this paper analyzes three path planning algorithms based on graph search,and proposes an improved D*Lite algorithm using the characteristics of minimum heap data structure.In addition,the attitude calculation and control algorithm of UAV is designed to ensure the flight stability of UAV.Finally,this paper builds a software and hardware platform.The simulation and experimental results show that the UAV path planning scheme based on stereo vision designed in this paper can realize the autonomous navigation of UAV.It has a certain reference value for the research in the field of autonomous navigation of UAV.