SAR与可见光图象匹配方法的研究

由于SAR和可见光成象传感器图象的物理特征不同，因此进行匹配难度很大。本文在深入分析了这两种传感器图象物理特征的基础上，提出了一种图象匹配的新方法。该方法运用形态学滤波方法去除SAR图象斑点噪声，利用图象的边缘特征进行匹配，并采用多分辨率分级搜索技术减少计算量，在对实验结果进行分析的基础上提出了搜索真实匹配位置的原则，大大提高了匹配的准确率。通过实验对SAR图象斑点噪声滤波和SAR与可见光图象匹配算法进行了验证，实验证明本算法的SAR图象去噪效果明显优于传统的Frost方法，图象匹配精度高，稳定性好。

Research on the Approach to SAR and Optical Images Matching

It is very difficult to match SAR(Synthetic Aperture Radar) and optical images because of the difference physical peculiarity. A new image matching approach based on analyzing the physical peculiarity of the two sensors has been proposed. This method applies morphological filtering algorithm to remove the speckle noise of SAR image and uses the edge's features of images to solve the matching problem and uses the technology of multiresolution gradational searching to reduce the complexity of computation. A new criterion that can search the real match position based on analyzing the experimental result is presented. It will be able to give a more accurate matching result. We validate the speckle noise filtering algorithm and the matching algorithm between the SAR image and Optical image by experiment. Experimental results show that the filtering effect is more excellent then traditional Frost algorithm, the matching algorithm have high precision and stability.