Software automation impact on aerial triangulation accuracy and geometric quality of orthophoto mosaic based on UltraCAM D sample images

Abstract

The current developments in photogrammetry software enable to obtain more detailed and higher resolution photoproducts in a shorter period of time. In addition to that, digital orthophoto production has become more automated, making the operator's task relatively easy. The present research investigates to which extent software automation affects the accuracy of aerial triangulation and geometric quality of orthophoto mosaic based on digital frame aerial camera UltraCAM D images. For this purpose, two separate projects were carried out: the first project examined the accuracy of aerial triangulation (AT) and the second one the geometric quality of the mosaic. The first project involved making two photo blocks: one with automatic and another with manual AT. In the second project, 3 orthophoto mosaics were made: one with fully automated processing steps for orthophoto production, the second with semi-automated steps with automatic AT and DEM correction, and the third one with manual steps. The Mean Root Square Error (RMS) quality of both projects was analyzed using ground control points. Both photo blocks of the first project didn't exceed the permitted size (RMS ≤ 0.45 m). The geometric accuracy of the manual AT was 0.094 m and that of the automatic AT was 0.415 m. The analysis of the geometric quality of the mosaics (second project) showed that geometrically the most accurate is the mosaic made with manual processing steps. The RMS was 0.308 m. The RMS of the semiautomatic mosaic was 0.335 m. The full-automatic mosaic was not within the permitted size: its RMS was 1.805.