Target geo-localization based on camera vision simulation of UAV

摘要

This paper presents a simulation study on estimating the Geo-Location of a target based on multiple image of the target taken from a gimbaled camera mounted on a unmanned aerial vehicle (UAV), which orbits around the target with a radius such that the target is always in the field of camera vision. The Camera Vision Simulation of the UAV is implemented by using an ortho Geo-TIFF (Geo-Spatial Tagged Information File Format) as imagery reference, positional and attitude attributes of UAV, Gimbal and Camera and internal characteristic of the simulated Camera. Target is localized using the simulation images taken from multiple bearing waypoints by applying the Geo-Location algorithm using the simulation parameters as reference. For improving the accuracy of the estimation, error reduction techniques like true average, moving average and recursive least square are also suggested and implemented.