A wavelet-based image matching scheme

Among many image matching methods, Hausdorff distance scheme is the most effective one. On the basis of Hausdorff distance scheme, we present Hausdorff-wavelet image matching algorithm in this paper. First, the method makes full use of wavelet transform modulus maxima to select interesting point which has magnitude and orientation information, and makes position and orientation of image or object certain. Then matches the images via Hausdorff distance scheme. Under the same precision, this new method can speed up the process by reducing large amount of unnecessary operation. The new matching-image method can detect translation, and rotation, and shows excellent robustness against random noise and salt noise. Key words: image matching, wavelet transform, wavelet transform modulus maxima, interesting point, Hausdorf