The Study of UAV-Photogrammetric Accuracy for Mapping Based on Variation of Ground Control Points

Abstract

This study focuses on the use of Unmanned Aerial Vehicle (UAV) photogrammetry for assessing the influence of distribution and number of Ground Control Points (GCP) that affect the   accuracy of mapping. The study site is final disposal site, which located in Nakorn Pathom province, 33 rai (5.28 ha). This study specified total of the GCP and Check Points (CPs) at 37 points, Ground Sampling Distance (GSD) at 5 cm/pixel, and flight configurations of 80% - frontal overlap and 75% - side overlap. The Agisoft PhotoScan Professional V.1.4.4 program was used for generating densified point cloud, digital elevation model, and orthoimage. The distribution formats of GCP were set to Edge distribution, Central distribution, Corner distribution, Half distribution, X-distribution, and Grid distribution. The best accuracy for horizontal and vertical positions were obtained from Grid distribution format. After that, this research studied the appropriate amount of GCP at 4, 6, 8, 10, 12, 14, and 16, respectively. The results show increased accuracy in both horizontal and vertical direction as the number of GCP increases and begins to stabilize when there are more than 8 GCP. According to the ASPRS 2014 standard, the obtained digital orthophoto can be used for the highest accuracy work. The vertical accuracy is defined at 5 cm-vertical accuracy class.