Small harbor detection in polarimetric SAR images based on coastline feature point merging
LIU Chun, YIN Junjun, YANG Jian
Department of Electronic Engineering, Tsinghua University, Beijing 100084, China
Abstract: A method was developed to automatic detect small harbors in polarimetric synthetic aperture radar (polSAR) images using coastline feature point merging based on analyses of structural characteristics of harbors. The coastline is accurately extracted by level set segmentation algorithm of polSAR with the coastline feature points then detected with a split and merge algorithm for digital curves. Then, the algorithm takes advantage of the characteristic that feature points along small harbor contour are denser than those along other coastline contours using a merging algorithm to detect the small harbors. The detection scheme was tested using polarimetric SAR images acquired by RADARSAT-2 over Singapore and the Zhanjiang area of China. The results show that almost all the harbors along the coastline are correctly detected by this method.
Key words: synthetic aperture radar    harbor detection    polarization    level set segmentation    point merge

1 基于水平集的海陆分割算法

 $\begin{array}{l} \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;E\left( {\Gamma ,\left\{ {{\Omega _1},{\Omega _2}} \right\}} \right) = \\ \nu \left| \Gamma \right| - \int\limits_{{\Omega _1}} {\ln f\left( {x|{\Omega _1}} \right){\rm{d}}x - } \int\limits_{{\Omega _2}} {\ln f\left( {x|{\Omega _2}} \right){\rm{d}}x.} \end{array}$ (1)

 $\hat \Gamma = \mathop {\min }\limits_\Gamma E\left( {\Gamma ,\left\{ {{\Omega _1},{\Omega _2}} \right\}} \right).$ (2)

 $\begin{array}{l} \;\;\;E\left( \Phi \right) = \nu \int\limits_\Omega {\left| {\nabla H\left( \Phi \right)} \right|} {\rm{d}}x - \\ \;\;\;\int\limits_\Omega {\left( {H\left( \Phi \right)\ln \left( {f\left( {x|{\Omega _1}} \right)} \right) + } \right.} \\ \left. {\left( {1 - H\left( \Phi \right)} \right)\ln f\left( {x|{\Omega _2}} \right)} \right){\rm{d}}x. \end{array}$ (3)

 $\frac{{\partial \Phi }}{{\partial t}} = - \delta \left( \Phi \right)\left( {\nu \kappa + \ln \frac{{f\left( {x|{\Omega _2}} \right)}}{{f\left( {x|{\Omega _1}} \right)}}} \right).$ (4)

 $\begin{array}{*{20}{c}} {f\left( {C|\Sigma ,L,p} \right) = }\\ {\frac{{{L^{pL}}{{\left( {\left| C \right|} \right)}^{L - p}}\exp \left( { - L{\rm{tr}}\left( {{\Sigma ^{ - 1}}C} \right)} \right)}}{{K\left( {L,p} \right){{\left( {\left| \Sigma \right|} \right)}^L}}}.} \end{array}$ (5)

 $\begin{array}{*{20}{c}} {\frac{{\partial \Phi }}{{\partial t}} = - \delta \left( \Phi \right) \cdot }\\ {\left( {\nu \kappa + L\left( {\ln \left| {{\Sigma _1}} \right| + {\rm{tr}}\left( {{\Sigma _1}^{ - 1}C} \right)} \right) - } \right.}\\ {\left. {L\left( {\ln \left| {{\Sigma _2}} \right| + {\rm{tr}}\left( {{\Sigma _2}^{ - 1}C} \right)} \right)} \right).} \end{array}$ (6)

2 基于岸线特征点合并的港口检测算法

 图 1 点对点的特征点合并算法

 图 2 闭合海岸轮廓线起始与终止特征点距离过近情况

3 实验结果

 图 3 新加坡部分海岸区域港口检测结果

 图 4 湛江部分海岸区域港口检测结果

4 结 论

 [1] 陈琪. SAR图像港口目标提取方法研究 [D].长沙:国防科学技术大学, 2011.CHEN Qi. Harbor extraction from SAR imagery [D]. Changsha: NUDT, 2011. (in Chinese) [2] 李艳,彭嘉雄.港口目标特征提取与识别 [J].华中科技大学学报:自然科学版, 2001, 29(6): 9-11.LI Yan, PENG Jiaxiong. Feature extraction of the harbor target and its recognition [J]. Journal of Huazhong University of Science and Technology, 2001, 29(6): 9-11. (in Chinese) [3] 陈琪,陆军,赵凌君,等.基于特征的 SAR遥感图像港口检测方法 [J].电子与信息学报, 2010, 32(6): 2873-2878.CHEN Qi, LU Jun, ZHAO Lingjun, et al. Harbor detection method of SAR remote sensing images based on feature [J]. Journal of Electronics & information technology, 2010, 32(6): 2873-2878. (in Chinese) [4] 张志龙,张焱,沈振康.基于特征谱的高分辨率遥感图像港口识别方法 [J].电子学报, 2010, 38(9): 2184-2188.ZHANG Zhilong, ZHANG Yan, SHEN Zhenkang. Port recognition in high resolution remote sensing images based on feature spectrum [J]. Chinese Journal of Electronics, 2010, 38(9): 2184-2188. (in Chinese) [5] 邢坤,付宜利.基于内港区域的港口目标识别 [J].电子与信息学报, 2009, 31(6): 1275-1278.XING Kun, FU Yili. Harbor target recognition based on inside region [J]. Journal of Electronics & information technology, 2009, 31(6): 1275-1278. (in Chinese) [6] Douglas D H, Peucker T K. Algorithms for the reduction of the number of points required to represent a digitized line or its caricature [J]. Cartographica: The International Journal for Geographic Information and Geovisualization, 1973, 10(2): 112-122. [7] Mason D C, Davenport I J. Accurate and efficient determination of the shoreline in ERS-1 SAR images [J]. IEEE Transactions on Geoscience and Remote Sensing, 1996, 34(5): 1243-1253. [8] Niedermeier A, Romaneessen E, Lehner S. Detection of coastlines in SAR images using wavelet methods [J]. IEEE Transactions on Geoscience and Remote Sensing, 2000, 38(5): 2270-2281. [9] Ben Ayed I, Mitiche A, Belhadj Z. Polarimetric image segmentation via maximum-likelihood approximation and efficient multiphase level-sets [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(9): 1493-1500. [10] Sonka M, Hlavac V, Boyle R. Image processing, analysis, and machine vision [M]. Toronto, Canada: Cengage Learning, 2014. [11] Rosenfeld A, Johnston E. Angle detection on digital curves [J]. IEEE Transactions on Computers, 1973, 100(9): 875-878. [12] Rosenfeld A, Weszka J S. An improved method of angle detection on digital curves [J]. IEEE Transactions on Computers, 1975, 24(9): 940-941.