An image registration method is proposed. A low altitude remote sensing image mosaicking system is development by using OpenCV library. We use the scale invariant feature transform (SIFT) points as the feature vector for image mosaicking. Local extreme points in scale space of the image are calculated. The SIFT feature points are extracted. The feature points' rough matching is performed based on the vectors Euclidean distance. We use random sample consensus (RANSAC) algorithm to optimize feature points matching. The image transform matrix is accurately estimated. Image fusion of overlap region is realized. The experimental results show that this method can solve the image registration problems which often occur in remote sensing images such as translation, scaling and rotation transform. Good mosaic results are achieved using the proposed method.