为解决合成孔径雷达（SAR）和光学图像之间由于非线性辐射和几何差异导致的匹配困难问题，提出一种新的基于特征的SAR和光学图像匹配算法。主要分为3个部分：特征提取、特征描述和特征匹配。在特征提取阶段，针对两者间的非线性辐射差异，使用基于相位一致性的方法得到SAR和光学图像的最小矩图和最大矩图，通过极值点检测与非极大值抑制提取出分布均匀且位置对应性好的最小矩点和最大矩点；在特征描述阶段，针对两者间的非线性几何差异，提出一种RS-GLOH描述子构建方法，主要使用斑点噪声抑制效果好的ROEWA算子和Sobel算子分别计算SAR和光学图像的梯度幅值和方向信息，然后建立梯度定位和方向直方图对特征点进行描述；在特征匹配阶段，结合最近邻距离比和快速抽样一致性分别匹配最小矩点和最大矩点。实验结果表明，本文方法具有旋转和尺度不变性，并且相比于PSO-SIFT、SAR-SIFT和OS-SIFT方法，本文方法在5组不同区域的SAR和光学图像对中可以得到更多的正确匹配点对数和更低的均方根误差。
In order to solve the difficulty in matching between synthetic aperture radar (SAR) and optical images due to nonlinear radiation differences and geometric differences, a new feature-based matching algorithm for SAR and optical images is proposed in this paper. This algorithm is mainly divided into three stages:feature extraction, feature description and feature matching. In the feature extraction stage, according to the nonlinear radiation difference between the two kinds of image, the minimum and maximum moment maps of SAR and optical images are obtained using the method based on phase consistency, and the minimum moment point and maximum moment point are obtained through extreme point detection and non-maximum suppression. In the feature description stage, in view of the nonlinear geometric difference between the two images, the ROEWA operator with good speckle noise suppression effects and Sobel operator are used to calculate the gradient amplitude and direction information of the SAR and optical images, respectively. Then gradient location orientation histogram (GLOH) is set up to describe the feature points (RS-GLOH). In the feature matching stage, method of combining the nearest neighbor matching (NNDR) and fast sampling consistency (FSC) is used to match the minimum and maximum moments respectively point. Experimental results show that the proposed method has excellent rotation and scale invariance. Compared with PSO-SIFT, SAR-SIFT, and OS-SIFT, more correct matching points and lower root mean square error (RMSE) can be obtained in 5 sets of SAR and optical image pairs in different regions.
[1] 李加元. 鲁棒性遥感影像特征匹配关键问题研究[D]. 武汉:武汉大学, 2018.
[2] 迟英朋, 刘畅. 一种适用于SAR图像配准的改进SIFT算法[J]. 中国科学院大学学报, 2019, 36(2):259-266. DOI:10.7523/j.issn.2095-6134.2019.02.014.
[3] Sui H G, Xu C, Liu J Y, et al. Automatic optical-to-SAR image registration by iterative line extraction and voronoi integrated spectral point matching[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(11):6058-6072. DOI:10.1109/TGRS.2015.2431498.
[4] Xiang Y M, Tao R S, Wang F, et al. Automatic registration of optical and SAR images via improved phase congruency model[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13:5847-5861. DOI:10.1109/JSTARS.2020.3026162.
[5] 张龙,肖俊,程晓龙,等. 基于多类型几何图元匹配的三维点云配准[J]. 中国科学院大学学报,2023,40(2):258-267. DOI:10.7523/j.ucas. 2021. 0047.
[6] Mellor M, Brady M. Phase mutual information as a similarity measure for registration[J]. Medical Image Analysis, 2005, 9(4):330-343. DOI:10.1016/j.media.2005.01.002.
[7] 尤红建, 胡岩峰. SAR和光学图像精配准技术的研究[J]. 雷达学报, 2014, 3(1):78-84. DOI:10.3724/SP.J.1300.2014.13154.
[8] Ye Y X, Shan J, Bruzzone L, et al. Robust registration of multimodal remote sensing images based on structural similarity[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(5):2941-2958. DOI:10.1109/TGRS.2017.2656380.
[9] Ye Y X, Wang M M, Hao S Y, et al. A novel keypoint detector combining corners and blobs for remote sensing image registration[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 18(3):451-455. DOI:10.1109/LGRS.2020.2980620.
[10] Li H, Manjunath B S, Mitra S K. A contour-based approach to multisensor image registration[J]. IEEE Transactions on Image Processing, 1995, 4(3):320-334. DOI:10.1109/83.366480.
[11] Yu Q Z, Ni D W, Jiang Y X, et al. Universal SAR and optical image registration via a novel SIFT framework based on nonlinear diffusion and a polar spatial-frequency descriptor[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 171:1-17. DOI:10.1016/j.isprsjprs.2020.10.019.
[12] Lowe D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2):91-110. DOI:10.1023/b:visi.0000029664.99615.94.
[13] Ma W P, Wen Z L, Wu Y, et al. Remote sensing image registration with modified SIFT and enhanced feature matching[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(1):3-7. DOI:10.1109/LGRS.2016.2600858.
[14] Dellinger F, Delon J, Gousseau Y, et al. SAR-SIFT:a SIFT-like algorithm for SAR images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(1):453-466. DOI:10.1109/TGRS.2014.2323552.
[15] Xiang Y M, Wang F, You H J. OS-SIFT:a robust SIFT-like algorithm for high-resolution optical-to-SAR image registration in suburban areas[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 56(6):3078-3090. DOI:10.1109/TGRS.2018.2790483.
[16] 李培, 姜刚, 马千里, 等. 结合张量与互信息的混合模型多模态图像配准方法[J]. 测绘学报, 2021, 50(7):916-929.
[17] Kovesi P. Phase congruency:a low-level image invariant[J]. Psychological Research, 2000, 64(2):136-148. DOI:10.1007/s004260000024.
[18] Kovesi P. Phase congruency detects corners and Edges[C]. Proceedings of the International Conference on Digital Image Computing:Techniques and Applications, Macquarie University, Sydney, Australia:DICTA, 2003:309-318.
[19] Fan J W, Wu Y, Li M, et al. SAR and optical image registration using nonlinear diffusion and phase congruency structural descriptor[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 56(9):5368-5379. DOI:10.1109/TGRS.2018.2815523.
[20] Li J, Hu Q, Ai M. RIFT:multi-modal image matching based on radiation-variation insensitive feature transform[J]. IEEE Transactions on Image Processing:a Publication of the IEEE Signal Processing Society, 2019:2019Dec17. DOI:10.1109/TIP.2019.2959244.
[21] Xie Z H, Liu J H, Liu C L, et al. Optical and SAR image registration using complexity analysis and binary descriptor in suburban areas[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19:1-5. DOI:10.1109/LGRS.2021. 3071870.
[22] Mikolajczyk K, Schmid C. A performance evaluation of local descriptors[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(10):1615-1630. DOI:10.1109/TPAMI.2005.188.
[23] Morrone M C, Owens R A. Feature detection from local energy[J]. Pattern Recognition Letters, 1987, 6(5):303-313. DOI:10.1016/0167-8655(87)90013-4.
[24] Fjortoft R, Lopes A, Marthon P, et al. An optimal multiedge detector for SAR image segmentation[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(3):793-802. DOI:10.1109/36.673672.
[25] Zhang X T, Wang Y H, Liu H W. Robust optical and SAR image registration based on OS-SIFT and cascaded sample consensus[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19:1-5. DOI:10.1109/LGRS.2021. 3069761.
[26] Hossain M T, Lv G H, Teng S W, et al. Improved symmetric-SIFT for multi-modal image registration[C]//2011 International Conference on Digital Image Computing:Techniques and Applications. December 6-8, 2011, Noosa, QLD, Australia. IEEE, 2011:197-202. DOI:10.1109/DICTA.2011.40.
[27] Schnabel R, Wahl R, Klein R. Efficient RANSAC for point-cloud shape detection[J]. Computer Graphics Forum, 2007, 26(2):214-226. DOI:10.1111/j.1467-8659.2007.01016.x.
[28] Wu Y, Ma W P, Gong M G, et al. A novel point-matching algorithm based on fast sample consensus for image registration[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(1):43-47. DOI:10.1109/LGRS.2014.2325970.
