Obstacle map building from USV using monocular camera

Abstract

In the modern world, technologies advance rapidly and can be found almost everywhere in our daily life. Unmanned Surface Vehicles (USV) is expected to play a vital role in the future military and civilian affair. In Singapore, the port has played a vital role in shaping the country’s economy and Singapore’s maritime industry. Thus, there is need to use USV to strengthen Singapore waters. In order for USV to accomplish harbor surveillance, it will be practical for USV to include situational awareness by sensing the immediate environment for locating obstacle. Therefore, in this report, we will be discussing the algorithm for image segmentation and obstacle detection for USV to have a situational awareness. We will use Linux Ubuntu version 16.04, OpenCV function and Robot Operating System (ROS) Kinetic as our operating system. The objective of this project is to assist in building a portion of a dynamic obstacle map from a monocular camera.