Combining Binocular Vision and Petri Net In Robot Path Planning

• Main Authors: Tsai, Yu-Ming, 蔡育銘
• Other Authors: Tsai, Lian-Jou
• Format: Others
• Language: zh-TW
• Published: 101
• Online Access: http://ndltd.ncl.edu.tw/handle/73813472140525345234

碩士 === 南台科技大學 === 電機工程系 === 100 === This thesis presents a robot routing algorithm using binocular vision and Petri Net. We equip two cameras as a robot visual perception system. First, we get each camera calibration parameters using checkerboard correction board. Second, through the acquisition and matching feature points, the system obtains image information and calculates the coordinates of point-depth information for making decision. In an unknown environment, if a robot wants to head a destination of some coordinate, it should get image information via robot visions, i.e. to get data such as location, distance, obstacles, and so forth. It is the image data combined in the decision-making theory of Petri Net in robotics. It gets more accurate in robot visual perception system by combining binocular vision system and Petri Net theory. In this thesis, we set up an experiment of calibration accuracy to compare results with camera calibration method. In addition, we also propose a test environment using the method. Through these experiments, we prove that it is feasible and suitable for robot routing algorithm.