Abstract
This paper proposes a vision-based control method for a micro helicopter over a low contrast ground. The 3D position, the posture and noise information are derived by a parallel template matching technique with edge detections and Gaussian filters. The proposed method does not require any known markers or special settings.