基于多尺度多方向Gabor模板的光学/SAR景象匹配

合成孔径雷达(SAR)成像制导通常采用光学基准图和SAR实时图进行特征提取和景象匹配.提出了一种光学/SAR异类影像匹配方法,利用多尺度多方向Gabor模板提取图像的Gabor特征后进行特征匹配,首先对SAR图像进行方向Frost滤波预处理,然后分别计算光学图像和SAR图像的高斯梯度图像,再利用多尺度多方向二维Gabor滤波器模板分别对两幅高斯梯度图像进行特征提取,最后对两组特征矩阵进行归一化互相关匹配.该方法直接利用光学图像和SAR实时图进行景象匹配,实验表明,该异类影像匹配方法较其他传统方法具有较高的鲁棒性和准确性.

Optical/SAR scene matching based on two-dimensional multi-scale and multi-direction Gabor templates

Generally optical reference image and SAR real-time image are used for feature extraction and scene matching in SAR(Synthetic Aperture Radar)imaging guidance.The heterogenous optical/SAR image matching method proposed in this paper extracts and matches image Gabor features using two-dimensional multi-scale and multi-direction Gabor templates.Firstly,directional Frost filtering preprocessing on SAR image is performed.Secondly,the Gaussian gradient image of the optical and SAR image are calculated respectively.Then,features are extracted from the two Gaussian gradient images using multi-scale and multi-direction two-dimensional Gabor filter group respectively.Finally,NCC(normalized cross-correlation) matching on the two characteristic matrixes is performed.The proposed method implements scene matching using optical image and real-time SAR image directly.The experiment result shows that the heterogenous image matching method is more robust and accurate than other traditional methods.