Abstract

Maritime transportation continues to maintain the largest volume in the transportation sector, and this is increasing day-by-day. This means there is an ever-increasing demand for safe navigation. In this study, we propose a stereo vision system for detecting and tracking ships as a tool for increasing navigational safety. Ships are detected from obtained images by using an edge detection algorithm that uses a moving average to find high-intensity gradients. The detected points are clustered and the 3D location of a detected ship is calculated from a stereo image pair. The detected ship is then tracked for several tens of seconds to obtain its course. The result of our experiment indicates that the proposed stereo vision system is useful as a tool for safe navigation.

Key words: Safety navigation, ship detection, tracking, computer vision, stereo vision.