Evaluating ortho-photo production potentials based on UAV real-time geo-referencing points

Abstract

Studies using unmanned aerial vehicle (UAV) are being actively conducted to obtain and apply spatial information. Spatial information based on orthophoto can be analyzed only when positional accuracy of these photos is high. To solve this problem, survey techniques, such as virtual reference station–real time kinematic (VRS–RTK) are used to create a ground control point (GCP), and geo-referencing is performed on the images. However, orthophoto based on non-fiducial points are generated in scenarios where spatial information is urgently required or in physical environments where GCPs are unlikely to be surveyed. In this method, a location error occurs and reduces the accuracy of orthophoto. In this study, a method that estimates GCPs using only measurement information using the UAV without a GCP survey and generates orthophoto based on those points, is proposed and its applicability is verified. 254 aerial photos were obtained using UAV. The location of six GCPs was estimated based on measurement information using UAV. Subsequently, the aerial photos and GCPs estimated based on the UAV measurement information were used to generate an orthophoto. This orthophoto was compared with those based on VRS–RTK GCPs and non-fiducial points in an error analysis. The analytic result indicates that the orthophoto based on the proposed method has an error that is 74.53% lower compared to that based on non-fiducial points. It is anticipated that the method proposed in this study to generate orthophoto based on information surveyed using the UAV can be effectively applied under physical environments where GCPs are unlikely to be established.