中文摘要：

由于成像条件变化造成的遥感图像之间的几何形变和灰度差异给影像匹配带来了困难，深入研究了SIFT特征描述符的生成方法，针对SIFT特征维数过高的问题进行改进，利用特征点邻域的圆形区域构造新的描述符，增强了描述符自身的抗旋转性，并降低了特征描述符的维数。实验表明，改进的特征描述符是可行有效的，在遥感影像目标匹配中取得满意的实验结果。

英文摘要：

Geometry deformation and grayscale distortion between remote sensing images caused by differences of imaging conditions are big problems in image matching. A deep research on SIFT descriptor creation method was made in this paper. Aiming at the problem of SIFT high-dimension descriptor, a new SIFT descriptor creation method was proposed, which took advantage of a circle region around the feature point to create the descriptor, and it could enhance rotation-resistance and reduce descriptor dimension. The experiments proved that the improved descriptor was effective and could get satisfied results in object matching of remote sensing images.