Vision-based target detection and path planning for autonomous navigation of underwater robots

Abstract

Underwater robots have been an emerging research area at the intersection of the field of robotics and oceanic engineering. Their applications include environmental monitoring, oceanographic mapping, and infrastructure inspections in deep sea. Performing these tasks, the ability of autonomous navigation is the key to a success, especially with the limited communications in underwater environments. Considering the highly dynamic and three dimensional environments, the autonomous navigation technologies including path planning and tracking have been one of the interesting but challenging tasks in the field of study. Cameras have not been at the center of attention as an underwater sensor due to the limited detection ranges and the poor visibility. Use of visual data from cameras is still an attractive method for underwater sensing and it is especially effective in the close range detections. In this thesis, the vision-based navigation techniques for underwater robots have been studied in depth. In order to overcome the limitations of cameras and to make use of the full advantages of image data, a number of approaches have been taken. The topics include color restoration algorithm for the degraded underwater images, detection and tracking methods for underwater targets, and robot guidance algorithms based on the visual detection results. The feasibilities of the proposed algorithms have been shown in the experiments with an underwater robot platform and the results have been analyzed both qualitatively and quantitatively.