Shoreline Detection in Images for Autonomous Boat Navigation

Our research is concerned with the analysis of images to assist in autonomous navigation of a boat. The problem is to determine the area of navigable region in an aquatic environment, by identifying the water areas or shoreline in the images. The problem is non-trivial as the image of an aquatic scene is highly unstructured, dynamic and changes significantly due to lighting, shadow and reflections. Therefore, it is challenging to identify the water boundary in a scene from an image. In this paper, we present a method developed for detecting the shoreline in an omni-directional image. We tested our method with various images from a river, lake and pond. The results show that our method correctly identified the shoreline in images and hence validating our approach.