目的相干斑的存在严重影响了极化合成孔径雷达(Pol SAR)的影像质量。对相干斑的抑制是使用SAR数据的必不可少的预处理程序。提出一种基于非局部加权的线性最小均方误差(LMMSE)滤波器的极化SAR滤波的方法。方法该方法的主要过程是利用非...目的相干斑的存在严重影响了极化合成孔径雷达(Pol SAR)的影像质量。对相干斑的抑制是使用SAR数据的必不可少的预处理程序。提出一种基于非局部加权的线性最小均方误差(LMMSE)滤波器的极化SAR滤波的方法。方法该方法的主要过程是利用非局部均值的理论来获取LMMSE估计器中像素样本的权重。同时,在样本像素的选取过程中,利用待处理像素的极化散射特性和邻域块的异质性来排除不相似像素以加速算法,同时达到保持点目标和自适应调节块窗口大小的目的。结果模拟影像和真实影像上进行的实验结果表明,采用这种方法滤波后影像的质量得到明显改善。和传统的LMMSE算法相比,无论是单视的影像还是多视的影像,本文方法去噪结果的等效视数都高出8视以上;峰值信噪比也提升了5.8 d B。同时,去噪后影像分类的总体精度也达到了83%以上,该方法的运行效率也比非局部均值算法有了较大提升。结论本文方法不仅能够有效抑制相干斑噪声,还能较好地保持边缘和细节信息以及极化散射特性。这将会为后续高效利用SAR数据提供保障。展开更多
The cryosphere is the frozen part of the Earth’s system. Snow and ice are the main constituents of the cryosphere and may be found in different states, such as snow, freshwater ice, sea ice, perma-frost, and continen...The cryosphere is the frozen part of the Earth’s system. Snow and ice are the main constituents of the cryosphere and may be found in different states, such as snow, freshwater ice, sea ice, perma-frost, and continental ice masses in the form of glaciers and ice sheets. The present review mainly deals with state-of-the-art applications of synthetic aperture radar (SAR) with a special emphasize on cryospheric information extraction. SAR is the most important active microwave remote sensing (RS) instrument for ice monitoring, which provides high-resolution images of the Earth’s surface. SAR is an ideal sensor in RS technology, which works in all-weather and day and night conditions to provide useful unprecedented information, especially in the cryospheric regions which are almost inaccessible areas on Earth. This paper addresses the technological evolution of SAR and its applications in studying the various components of the cryosphere. The arrival of SAR radically changed the capabilities of information extraction related to ice type, new ice formation, and ice thickness. SAR applications can be divided into two broad classes-polarimetric applications and interferometric applications. Polarimetric SAR has been effectively used for mapping calving fronts, crevasses, surface structures, sea ice, detection of icebergs, etc. The paper also summarizes both the operational and climate change research by using SAR for sea ice parameter detection. Digital elevation model (DEM) generation and glacier velocity mapping are the two most important applications used in cryosphere using SAR interferometry or interferometric SAR (InSAR). Space-borne InSAR techniques for measuring ice flow velocity and topography have developed rapidly over the last decade. InSAR is capable of measuring ice motion that has radically changed the science of glaciers and ice sheets. Measurement of temperate glacier velocities and surface characteristics by using airborne and space-borne interferometric satellite images have been the significant a展开更多
文摘目的相干斑的存在严重影响了极化合成孔径雷达(Pol SAR)的影像质量。对相干斑的抑制是使用SAR数据的必不可少的预处理程序。提出一种基于非局部加权的线性最小均方误差(LMMSE)滤波器的极化SAR滤波的方法。方法该方法的主要过程是利用非局部均值的理论来获取LMMSE估计器中像素样本的权重。同时,在样本像素的选取过程中,利用待处理像素的极化散射特性和邻域块的异质性来排除不相似像素以加速算法,同时达到保持点目标和自适应调节块窗口大小的目的。结果模拟影像和真实影像上进行的实验结果表明,采用这种方法滤波后影像的质量得到明显改善。和传统的LMMSE算法相比,无论是单视的影像还是多视的影像,本文方法去噪结果的等效视数都高出8视以上;峰值信噪比也提升了5.8 d B。同时,去噪后影像分类的总体精度也达到了83%以上,该方法的运行效率也比非局部均值算法有了较大提升。结论本文方法不仅能够有效抑制相干斑噪声,还能较好地保持边缘和细节信息以及极化散射特性。这将会为后续高效利用SAR数据提供保障。
文摘为了充分利用极化合成孔径雷达(synthetic apeture radar,SAR)图像丰富的地物信息并解决单一特征在图像分类中的局限性问题,提出了一种基于特征选择双层支持向量机(support vector machine,SVM)的特征融合算法,充分利用特征间的完备性和互异性,以形成更有效的特征组合,并用于SAR图像的分类。首先,对SAR图像进行多种类型特征矢量的提取以能完整地描述全极化SAR图像;其次,进行特征归一化处理,以保证不同的特征向量在同一准则下进行选择,以期在进行分类时具有相同的作用;再次,引入空间金字塔(spatial pyramid,SP)分块提取不同尺度的特征矢量;然后,利用最小冗余最大关联(minimum redundancy and maximum relevance,mRMR)特征选择方法获取每种类别的最优特征子集,避免各类特征的简单组合导致的特征冗余和过度拟合现象;最后,引入多层的思想,构造双层SVM模型,实现单层目标类别概率的优化和再处理。实验结果验证了该算法对于极化SAR图像分类的有效性。
文摘The cryosphere is the frozen part of the Earth’s system. Snow and ice are the main constituents of the cryosphere and may be found in different states, such as snow, freshwater ice, sea ice, perma-frost, and continental ice masses in the form of glaciers and ice sheets. The present review mainly deals with state-of-the-art applications of synthetic aperture radar (SAR) with a special emphasize on cryospheric information extraction. SAR is the most important active microwave remote sensing (RS) instrument for ice monitoring, which provides high-resolution images of the Earth’s surface. SAR is an ideal sensor in RS technology, which works in all-weather and day and night conditions to provide useful unprecedented information, especially in the cryospheric regions which are almost inaccessible areas on Earth. This paper addresses the technological evolution of SAR and its applications in studying the various components of the cryosphere. The arrival of SAR radically changed the capabilities of information extraction related to ice type, new ice formation, and ice thickness. SAR applications can be divided into two broad classes-polarimetric applications and interferometric applications. Polarimetric SAR has been effectively used for mapping calving fronts, crevasses, surface structures, sea ice, detection of icebergs, etc. The paper also summarizes both the operational and climate change research by using SAR for sea ice parameter detection. Digital elevation model (DEM) generation and glacier velocity mapping are the two most important applications used in cryosphere using SAR interferometry or interferometric SAR (InSAR). Space-borne InSAR techniques for measuring ice flow velocity and topography have developed rapidly over the last decade. InSAR is capable of measuring ice motion that has radically changed the science of glaciers and ice sheets. Measurement of temperate glacier velocities and surface characteristics by using airborne and space-borne interferometric satellite images have been the significant a
