High speed obstacle detection and tracking using stereo vision

• Main Author: Guo, Jasper Yonghui.
• Other Authors: Wang Han
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/50123
• Institution: Nanyang Technological University

This is the second year of an on-going industrial project on USV (Unmanned surface vessel), where Stereo vision is being used to detect obstacles in the open sea. However, the detection of obstacles is sometimes lost temporarily due to occlusion or overlapping (When a big obstacle overshadows a small one, the small obstacle is usually lost and absorbed into the bid one). In order to improve this, the student has chosen to use template matching technique for tracking to enhance the reliability of detection of obstacles so that once obstacles are detected, the obstacles remain detected even if occlusion occurs until the obstacles leaves the USV’s field of vision. Using template matching has improved the robustness overall system, however in situations where the current USV obstacle detection algorithm fails to detect an obstacle, template matching is unable to track the obstacle.