GEOMETRIC CORRECTION OF SMARTPHONE IMAGES USING UAVs REFERENCE IMAGES FOR FACILITY MAINTENANCE

摘要

Continuous facility maintenance is important to prevent accidents. Currently, the method of facility maintenance is dependent on visual inspection, which is labor-intensive, time-consuming and has a risk of casualties. In addition, the results of facility inspection can be different according to inspectors. In this paper, we propose an image-based inspection method, which measures location of points of interest using a smartphone image for effective facility maintenance. We performed georeferencing of smartphone images using pre-processed unmanned aerial vehicles (UAVs) reference images, not ground control points (GCPs). We then rectified the images using position of smartphone images derived from georeferencing onto a target plane of the facility. Experimental results show that the proposed approach has accuracies of 3.543 cm, 0.209° in the position of smartphone images and 9.829 cm in the location of points measured from images. We expect that this method can enhance utilization of image based facility maintenance.