基于特征提取的光学影像和SAR影像自动配准

针对现有的配准方法用于多光谱影像与SAR遥感影像配准时，存在受SAR图像斑纹噪声影像大、手工选取配准控制点精度低、利用图像景物特征配准时获取区域和边沿困难等问题，以SPOT5影像与RADARSATSAR影像配准进行实验，提出了一种利用改进的SIFT在提取的特征图像上寻找匹配点进行粗配准，然后利用交叉累积剩余熵作为相似性测度结合原始影像信息寻找光学特征图像的角点在SAR影像上的匹配点并进行精配准的方法，配准精度达到了子像素级水平。实验结果表明该方法对多源遥感影像有很强的适应性，配准精度高。

Automatic Registration for Optical and SAR Remote Sensing Images Based on Feature-extraction

In the registration for SAR and optical images, traditional methods are limited by SAR speckle noise and the accuracies of selected registration control points. Take automatic registration for optical and SAR remote sensing images for example. In order to registrate roughly, SIFT is appled to the characteristic images which have been got to get match points. And register SAR to optical image accurately with the help of Cross-Cumulative Residual Entropy; comers on the optical characteristic image and the original image information. And the registration accuracy get to sub-pixel. Experimental results show that the method can accomplish high accuracy and automatic registration for multi-sensor images.