Low Cost Technique of Enhancement Georeferencing for UAV Linear Projects

Abstract

Global positioning System (GPS) on Unmanned Aerial Vehicles (UAV) platform depend on MEMS (Micro Electro Mechanical Systems) technology with 10 m accuracy at UAV camera position at shooting time. However, obstacles to GPS signal at the finest flight altitude might avoid accurate retrieval of camera station positions. The aim of this investigation is getting accurate photogrammetric products in UAV linear projects by improving the georeferencing, using a little number of Ground Control Points (GCPs) in a small part and improved camera station positions. This technique is low cost and helpful in disasters especially in projects that cannot be distributing GCPs all over it. DJI PHANTOM 4 PRO acquired 390 images in three flight lines with average flight altitude about 40 m along 1000 m length of study area. The ground sample distance (GSD) is about 1.2 cm. The average end lap and side lap are about 80%. The dimension of area covered by this area is 1000 m by 90 m. Methodology of enhancement GPS auxiliary data is by using Linear Relation (LR) model. In LR model, correlation for the same specified images between the auxiliary by GPS on UAV platform and computed photo centers coordinates through georeferencing using a little of GCPs in small part of project has been determined. The specified images only contained GCPs, which are used in georeferencing. Moreover, using LR model to generate new coordinate values for the rest images to use them in georeferencing. The errors assessment was done using 86 checkpoints (CPs) not used in georeferencing distributed all over the study area. The using of the generated photo centers coordinates by LR model with the four GCPs gives accurate results in georeferencing better than using only the same four GCPs.