Research On Key Technologies Of UAV Visual Localization And Mapping In Low Illumination Environment

In recent years,as a new type of air mobile platform,UAV has developed rapidly,and its intelligent level has made rapid progress,especially in the field o f c ivil a nd m ilitary applications.Whether the UAV can get accurate position and posture information in complex full-time environment decides whether it can effectively accomplish its task independently.Therefore,the high-precision positioning and mapping of UAV is an important technology in the field of UAV research.In this paper,the passive UAV and the active UAV are proposed and constructed.Based on the research of vision positioning and mapping system,a full-time UAV positioning and mapping algorithm is proposed for low illumination environment.The main contents of the thesis are summarized as follows:(1)Firstly,the development of UAV autonomy technology is sorted out.On this basis,the key technology of UAV autonomy,high precision vision positioning and mapping,is introduced.The achievements of domestic and foreign scholars in UAV vision positioning and mapping technology are summarized,and two different positioning and mapping technologies are explained in detail.The application and development status will provide some reference for future research.(2)The key technologies of UAV positioning and mapping tasks are analyzed and summarized.Firstly,the geometric model of camera imaging set and the principle of stereo vision are introduced.Then,the method of UAV visual positioning and mapping is discussed in detail,which provides a theoretical basis for future research.(3)A rotor UAV positioning system based on panoramic camera array is proposed and constructed.The system solves the problem of locating the UAV in low altitude and GPS-free autonomous recovery and landing by using panoramic camera arrays arranged arbitrarily on the ground.Firstly,fish-eye c amera i s s elected t o o btain t he w ide-area o bservation fi eld of view,and a fast self-calibration algorithm of camera array is proposed for the imaging system.Then,the UAV is detected by the camera array of Lubang and the coordinated high-precision positioning task is further realized.The system has the advantages of quick layout and high positioning accuracy.(4)A vision localization and mapping algorithm for UAVs in low illumination environment is proposed.Aiming at the problem of image invisibility in low illumination,an algorithm of UAV visual positioning and mapping in low illumination environment is proposed by using near infrared camera and active light compensation of near infrared laser lamp,which effectively solves the problem of UAV positioning in low illumination environment.This algorithm is not only superior to ORB-SLAM and other open source algorithms in sufficient illumination environment,but also quite stable in low illumination environment.(5)This paper proposes an incremental UAV visual localization and mapping algorithm in full time.Through the innovation of hardware system and algorithm,the feature extraction and matching of images in different illumination environments are realized,and the incremental visual positioning and mapping problem of UAV in full-time environment is solved.Through a large number of simulation and real experiments,the robustness and effectiveness of the algorithm are verified effectively.