A CASE STUDY OF MULTI-HEAD CAMERA SYSTEMS ON UAV FOR THE GENERATION OF HIGH-QUALITY 3D MAPPING

Abstract

This research aims to present the geometrical model of camera-rig and numerical modeling of the “FOXTECH 3DM V3” multi-head camera system. The multi-head camera system consists of one nadir camera and four 45-degree oblique cameras. The final geometric model of the camera system should be consistent with physical camera geometry to confirm the correctness of the model. The research reveals the parameterization and numerical modeling of “3DM-V3” multi-head camera systems that all four sets of relative translation parameters between the reference camera and all four oblique cameras named , when  has a value from 1 to 4, have to be constraint of a nominal value. The other four sets of relative rotation parameters between the reference camera referred to the other four oblique cameras named , when  has a value from 1 to 4 could be initialized but set “free” during the further numerical process. Other conventional parameters representing all 5 internal camera models (f, ,R1,R2,R3,T1,T2) could be set “free” as widely practice and they will be numerically adjusted with success. The final camera model resulting from adjustment computation for triangulation affirms precise geometry with a mean reprojection error of 0.225 pixels and the rig parameters are consistent with the physical characteristics of the multi-head camera system. In this research, the triangulation block is tied up with ground constraints and check-points using14 signalized GCPs from Chulalongkorn UAV Test Field of size 0.8 square kilometers in Saraburi province. The final adjusted parameters for GCPs show RMSE 0.023 meter in horizontal and RMSE 0.019 meter for the vertical component.