Current Proceedings on Technology
Yazarlar: Mohsen. H. Tehrani, M. A. Garratt, S. Anavatti
Konular:-
Anahtar Kelimeler:Panoramic vision,Attitude determination,Least-Square Estimation,Random Sample Consensus,Kalman Filter
Özet: An attitude estimation based on panoramic vision is presented using a panoramic Ultraviolet (UV) vision system along with an Extended Kalman Filter. The pitch and the roll are estimated from vision processing in the presence of sun, clouds and occluding features. The vision system uses real-time vision processing and can be used independently to give an accurate attitude estimation regarding the position of the horizon. An Extended Kalman Filter is used as a sensor fusion technique to reduce the visual attitude estimation error that occurs when close objects are in the field of vision and to smoothen the approximation by integrating the visual and inertial sensors.