研究目的
To determine the optimum number of ground control points (GCPs) necessary for deriving high precision results from UAS images.
研究成果
The study concluded that a density of 1 GCP/200 m2 is necessary to obtain high accuracy of the final products from UAS images. The optimum number of GCPs for georeferencing the nadiral UAS images was found to be 14 for the 28 m flight and 15 for the 35 m flight when using Pix4D software, and 20 for the 28 m flight and 15 for the 35 m flight when using 3DF Zephyr Pro software. The results demonstrated a clear overview of the number of GCPs needed for the indirect georeferencing process with minimum influence on the final results.
研究不足
The study was limited by the specific conditions of the study area and the equipment used. The accuracy of the final products could be influenced by factors such as the camera's focal length, flight altitude, camera orientation, image quality, processing software, and the precision with which targets can be measured and matched.
1:Experimental Design and Method Selection:
The study involved photographing an area of about 1 ha with a DJI Phantom 3 Standard UAS at two different heights (28 m and 35 m above ground) and measuring 50 GCPs using a total station. The UAS images were processed using Pix4D Mapper Pro and 3DF Zephyr software with a full bundle adjustment process, gradually increasing the number of GCPs from three to 40. A third test was conducted using 3DF Zephyr Pro software with a free-network approach in the bundle adjustment. The point clouds and mesh surfaces derived from these processes were compared with a TLS point cloud for accuracy assessment.
2:A third test was conducted using 3DF Zephyr Pro software with a free-network approach in the bundle adjustment. The point clouds and mesh surfaces derived from these processes were compared with a TLS point cloud for accuracy assessment.
Sample Selection and Data Sources:
2. Sample Selection and Data Sources: The study area was located near the Faculty of Hydrotechnical Engineering, Geodesy and Environmental Engineering at the 'Gheorghe Asachi' Technical University of Iasi, Romania. The 50 GCPs were uniformly distributed over the study area.
3:List of Experimental Equipment and Materials:
DJI Phantom 3 Standard UAS, total station for measuring GCPs, Pix4D Mapper Pro software, 3DF Zephyr Pro software, Maptek I-Site 8820 terrestrial laser scanner.
4:Experimental Procedures and Operational Workflow:
The UAS images were processed with a minimum number of GCPs, gradually increasing to 40, with the remaining points serving as check points for accuracy assessment. The accuracy was evaluated by comparing the manually measured coordinates of the CPs on each oriented image with those determined by GNSS technology.
5:Data Analysis Methods:
The Euclidian distance between the two coordinate sets for a point was calculated, followed by the determination of the root mean square error (RMSE). The accuracy of the 3D point clouds and mesh surfaces was assessed by comparing them with the TLS point cloud using CloudCompare software.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容