中文摘要：

图像配准是计算机视觉中诸多问题的基础，基于图像特征的配准方法仍然是该领域的研究热门。为了提高算法的效率，拥有更好的实用性，提出了一种基于FAST-DAISY的遥感图像配准方法。首先运用FAST算法提取特征点，提出分配主方向的方法，利用DAISY算法建立描述符，得到特征点集后，使用RANSAC（随机抽样一致性）算法剔除误匹配点对，最终估计仿射变换参数，利用二次线性插值法得到配准后的遥感图像。实验结果表明，算法对于平移、旋转、灰度差异、地物差异、位置差异、小尺度差异和噪声干扰的遥感图像有较好的配准效果，匹配时间通常介于SIFT与SURF-DAISY算法之间，算法在实用性上有较大优势。

英文摘要：

Image registration is the basis of many problems in computer vision, where feature-based image registration methods have been widely used. In order to improve the efficiency and usability of the registration algorithm, this paper presented a re- mote sensing image registration method based on FAST-DAISY algorithm. Firstly, it used FAST algorithm to extract feature points and proposed a main orientation distribution method. Then, utilized DAISY algorithm to establish descriptors. After get- ting the feature point set,it used the RANSAC algorithm to eliminate wrong matches. Finally, estimated the affine transformation parameters, and accomplished the image registration process by using quadratic linear interpolation method. The experimental results indicate that the proposed method has a better registration result for the translation, rotation, gray-scale differences, fea- ture differences, position differences, small-scale differences and noise interferences of remote sensing images. The time com- plexity is usually between the SIFT and SURF-DAISY algorithms. Consequently, the proposed algorithm is suitable for practical applications.