Georeferencing and rectification of smartphone images using UAV reference images for monitoring infrastructure close to a river

The time and cost for recovering damaged infrastructure close to a river after disaster like typhoon, heavy rain and earthquake are increasing. It is thus important to monitor the infrastructure for maintenance and damage assessment effectively, particularly using remote sensors such as cameras. If we utilize smartphone images for monitoring the infrastructure, we can save significantly the time and cost to acquire up-to-date images. In this study, we develop a method to process smartphone images for infrastructure monitoring, using accurately pre-processed Unmanned Aerial Vehicles (UAV) reference images. First, we perform georeferencing of smartphone images based on their correspondence with UAV reference images. Then we project these georeferenced images onto target object surfaces. In this process, we derive the camera position and orientation for each smartphone image and the 3D coordinates of main object points appearing in smartphone images. We then verified the results by comparing them to surveyed reference data. It shows that we can utilize the rectified smartphone images to find the locations, length of cracks and areas of damaged structures. The proposed approach shows that smartphone images can be successfully used to monitor the damaged infrastructure close to a river.