Robot control with graph based edge measure in real time image frames
This paper proposed a visual servoing system for autonomous mobile robot navigation under an eye-to-hand camera configuration using omnidirectional vision. The proposed algorithm is based upon Position-Based Visual Servoing (PBVS) and uses only one fixed camera to provide the global path planning. The image is captured by the omnidirectional camera mounted on the ceiling followed by camera calibration, visual pose detection of targets, obstacles and robot, and path planning. The visibility graphs and Dijkstra's algorithm are adopted to find shortest path. The presented method simplified the algorithm to reduce complexity for real-time operation. The experimental results demonstrated the effectiveness and efficiency of the proposed visual servoing algorithm.