high computational complexity in image registration method based on speed-up robust features (SURF) algorithm
an improved SURF algorithm is proposed.At first it uses SURF to extract the feature points from two images.The feature points with higher entropy are extracted.By using the distance threshold to reduce the number of feature points
the computation work can be reduced.PROSAC algorithm is used to reject false matches and compute homography matrix.The experiments verify the algorithms efficiency and robustness.