Comparative Study of UAV Mapping with and without Ground Control Points (GCPs)

Abstract

Unmanned Aerial Vehicles (UAVs) have transformed geospatial data acquisition, yet the necessity of Ground Control Points (GCPs) for spatial accuracy versus operational efficiency remains a critical topic of investigation. This study presents a comparative analysis of UAV mapping accuracies achieved with and without the integration of GCPs. Conducted at the Asian Institute of Technology (AIT) in Pathum Thani, Thailand, the research utilized a DJI Phantom 4 Pro for aerial imagery and a Javad TRIUMPH-2 GNSS receiver to establish survey-grade GCPs and Check Points (CPs) via Real-Time Kinematic (RTK) methods. Flight parameters included a 90-meter altitude with 70% forward and side overlap, yielding a Ground Sampling Distance (GSD) of 3.45 cm/pixel across a 0.252 km² area. The photogrammetric processing resulted in a Root Mean Square Error (RMSE) of 26.26 cm for GCPs and 76.76 cm for independent CPs. The findings indicate that while orthophotos and Digital Surface Models (DSMs) processed without GCPs exhibit smoother visual aesthetics and slightly more visible detail, they suffer from significant systematic vertical errors, such as the "doming" effect, and lack absolute geospatial accuracy. Conversely, GCP-constrained models achieve much higher absolute positional accuracy but reveal geometric distortions around vertical structures as the software forces the imagery to match precise ground coordinates. Ultimately, GCPs are essential for mapping projects requiring rigorous absolute accuracy, whereas unconstrained workflows may suffice for relative topographic visualization.