원문정보
초록
영어
The problem of estimating motions of a mobile robot has been studied extensively in the past. In this paper, a new Vision-Aided Motion Detection Algorithm (VAMD) for orientation sensing in an unknown environment is proposed. This algorithm aims to estimate a mobile vehicle orientation, based on a single camera pose, while reducing accumulation error. The VAMD makes use of feature points, detected in a 2D scene, to recognize any rotational motion of a camera, placed on a mobile platform, and estimate the orientations in 3 Degree of Freedom (pitch, yaw, roll angles) from its initial position. The robustness, accuracy and flexibility of this algorithm will be put through several experiments on real scenarios aided with mobile robot, and the results are compared with data obtained from the mobile robot optical encoder.
목차
1. Introduction
2. Vision-Based Motion Detection Algorithm (VAMD)
2.1. VAMD Flowchart and Framework
2.2. Orientation Estimation
3. Experimental St
3.1. Experimental Setup
3.2. Optical Flow Pattern over Rotational and Translational Motions
3.3. VAMD Accuracy
3.4. VAMD Robustness
4. Conclusion
Acknowledgements
References