【摘要】 提出了一种新的基于边缘的合成孔径雷达(SAR)图像与卫星光学图像匹配算法.在这种算法中,针对SAR图像固有特性提出了一种边缘特征的提取方法,并提出了一种新的基于边缘连续性密度加权的子模板Hausdorff距离矩阵度量方式,采用子模板Hausdorff距离矩阵的Frobenius范数来描述整个矩阵的"大小"的尺度,从而建立图像相似性测度.这种测度具有可靠,鲁棒,易于并行实现的优点.试验结果表明,这种算法鲁棒性好,匹配精度高,计算速度快.更多还原

【Abstract】 In this paper,a new edge-based synthetic aperture radar(SAR) image and satellite optical image matching algorithm has been put forward.In this algorithm,first a edge feature extraction method has been proposed for the inherent characteristics of SAR image,and a new sub-templates Hausdorff metric distance matrix method based on the edge continuity of density-weighted,using the Frobenius norm of sub-templates Hausdorff distance matrix to describe the scale of entire matrix size,establishing image similarity measurement.This measurement is reliable,robust,and has the advantages of easy parallel implementation.Test results show that this algorithm has good robustness,high matching precision,high calculation speed.更多还原