A vision based target detection and tracking system

• Main Author: Ma, Qing Qiang
• Other Authors: Xie Lihua
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics
• Online Access: http://hdl.handle.net/10356/40803
• Institution: Nanyang Technological University

Computer vision is a diverse and relatively new field of study. Some of the typical tasks of Computer vision are: object detection and recognition, tracking and motion analysis, scene reconstruction and image restoration (noise removal). This project utilized the knowledge of computer vision and C++ programming. In this project, a vision system was built which can detect and identify different robots from the images captured by up to four cameras and compute the position and orientation of each robot. The detection and identification of the robots were based on color detection, since the robots were specified by three colored papers at the top. The color setting (left and right side) of each robot was unique, thus the identification of the robots was achieved by examining the colors and their relative positions. Positions of the robots in terms of field coordinates were calculated using the coordinates in the image and a transformation matrix which was pre-calculated using some well positioned points. And the orientations of the robots were calculated by the relative positions of the different color papers also. Finally, the information of the robots calculated were transmitted to a LabView program which was running on another computer via wireless communication using TCP/IP. This thesis provides the detailed tasks of the project and the main theories and the methodologies used to achieve the tasks, also some limitations of the vision system and some recommendations are also included.