Detecting And Tracking Objects In An Autonomous Surface Vehicle Using Stereo Vision

Nowadays, one of the main challenges in the maritime is the control and monitoring of ports in different weather conditions. To achieve this, Autonomous Surface Vehicles (ASV) are used. In this research, as part of the Morvarid research project, a stereo vision system was implemented on an ASV to detect, localize, and track static and dynamic objects. In order to examine the proposed obstacle detection algorithm, two sets of experiments have been designed; first, the ASV moves toward stationary objects to localize all static and dynamic objects. Second, the ASV tracks a target boat, which equipped with an RTK-GPS and estimates relative and absolute positions and movement vector. The results showed that the proposed algorithm successfully estimated 3D size and position of the surrounding objects. The accuracy of the tracking algorithm was evaluated in the digital elevation map, with the lowest RMSE of 1.575 (m), and 0.6621 (m) using raw and EKF data, respectively.