SOME ASPECTS IN HEIGHT MEASUREMENT BY UAV PHOTOGRAMMETRY

Abstract. We conducted an experiment to investigate the feasibility of the deformation measurement of a large-scale solar power plant on reclaimed land by UAV photogrammetry. Two teams engaged in the experiment at first. One, which is called Team-A, carried out orientation of images following the procedure of conventional aerial photogrammetry. The other, which is called Team-C, executed that in the manner of close range photogrammetry. The RMSE in height measurement by Team-A was 121.5 mm, while that by Team-C was 8.7 mm. This paper reports an analysis conducted in order to investigate the cause of the large difference in height measurement accuracy between Team-A and Team-C. In the analysis the third team, which is called Team-S, conducts supplementary orientation by using the images utilized by Team-A in the same manner as Team-C did. The RMSE in height measurement by Team-S is 19.1 mm. Our investigation focuses on the difference of the arrangement of points utilized in the orientation. Team-A selected pass points and tie points on image automatically by Intergraph’s ImageStation Automatic Triangulation (ISAT) software, while Team-C and Team-S selected points to be utilized in orientation manually so that selected points are distributed uniformly in the experiment area. From the results of the analysis we conclude that the sets of tie points along a straight line on a plane that were selected automatically by the ISAT would bring the low accuracy in height measurement by Team-A.