ENHANCING PERFORMANCE AND EVALUATING ACCURACY OF GNSS POSITIONING FOR UAV MAPPING

Abstract

This research aims to improve the efficiency and accuracy of GNSS data obtained from Unmanned Aerial Vehicles (UAV) for mapping. The study compares the Root-Mean-Square Error (RMSE) on the horizontal and vertical components processed by photogrammetric Aerial Triangulation (AT) with Ground Control Point (GCP) used as checkpoints obtained from different processing methods and software. Five cases are examined namely: case 1, RTK on-board M300; case 2, PPK using scientific software; case 3, PPK using commercial software; case 4, PPK with precise clock and orbit; and case 5, k-PPP with precise clock and orbit neglecting the base station. To demonstrate the quality of the data obtained from each case, the research compares the difference of the same flight trajectory obtained from commonly used GNSS processing software for map production. The aerial images were recorded from DJI MATRICE 300 UAV equipped with a ZENMUSE P1 camera with a 45-megapixel image resolution. The test flight captured images over an area of approximately 1 square kilometer, with 49 ground control points used as checkpoints for processing aerial triangulation by PIX4Dmapper software. The study found that the best accuracy of check points was achieved in case 3, PPK using commercial software, with maximum uncertainty less than 5,9 centimeters. In term of comparing differences of flight trajectory by applying precise orbit and clock, they showed that the difference between case 3 and case 4 was insignificant. Additionally, the difference in trajectory between case 3, PPK using commercial software, and case 5, k-PPP with precise clock and orbit neglecting the base station, showed a systematic error of -0.4 meters in the east and 0.1 meters in the north for the horizontal component. The vertical difference was found to be accurate to the centimeter level compared to prior PPK processing.