Research And Application Of 3D Reconstruction Technology And Track Planning Based On Unmanned Aerial Vehicle

In recent year,unmanned aerial vehicle or drone has been developing rapidly in China and other countries.Fruitful results have been yielded in both military and civil application.On one hand,drone has shown us a future direction for military operations under the condition of informationization.It’s an advanced comprehensive weapons system that combining detection,identification,tacking,and deciding.On the other hand,with a higher demand of visualization of geographic and topographic information,the informationization of topography from 2D to 3D is gaining momentum.Thus,3D reconstruction technology has drawn people’s attention.Meanwhile,given the complexity of mission environment and the increase of uncertainties,it becomes crucial to manage drones effectively.Regarding management of drones,it is significant to plan the fight path properly to avoid barriers and fly along the shortest route to reach the destination.This thesis is about 3D reconstruction of aerial images and the planning of flight path of drones.Firstly,an analysis is conducted on the flight control system platform of drones that based on APM autopilot.And assembling of drone structure,building of ground equipment,designing of hardware modules,programming of software,setting of system parameters and communication simulation test are done by sequence.Then,acquire serial aerial images by remote sensing of drones,pick and pair of characteristic points by using method of SIFT,thus DEM data are generated.So that 3D reconstruction of topography in Open GL is realized.Finally,in order to get the drone adapted into complicated environment,simulation of topographic emulation algorithm,threatening peaks algorithm and equivalent topography fusion algorithm are built in function by study of the features of digital map.To optimize the flight path of the drone,an advanced PSO algorithm is adopted.With this,the researcher built basic constraint conditions during the flight and evaluation function of the flight path;built modeling of flight planning environment and evaluation system;introduced the concept of minimum threatening curved surface;thus,turning the problem of 3D flight path planning into the problem of 2D route optimizing.Under the guidance of the evaluation system,all flight paths are generated to meet the demands of varied flight path constraint conditions.Flight path smoothing algorithm is presented to simulate the actual problem in flight path planning utilizing MATLAB.It is shown that this designed drone owns stable system and is in good flight condition.The reconstruction algorithm that bases on aerial images could meet the demands of 3D visualization.In flight path planning,the system simulates targeting the shortest flight path while meeting varied constraint conditions.As in the results,the improved particle swarm algorithm is more efficient than the basic particle swarm algorithm,for the former one could generate a better flight path under constraint conditions.It can be realized and applied into flight path planning of drones with a promising future of utilization.