Multipolarization Radar Images for Geologic Mapping and Vegetation Discrimination

The NASA/JPL airborne synthetic aperture radar system produces radar image data simultaneously in four linear polarizations (HH, VV, VH, HV) at 24.6-cm wavelength (L-band), with 10-m resolution, across a swath width of approximately 10 km. The signal data are recorded optically and digitally and annotated in each of the channels to facilitate a completely automated digital correlation. Both standard amplitude, and also phase difference images are produced in the correlation process. Individual polarization and range-dependent gain functions improve the effective dynamic range, but as yet do not permit absolute quantitative measurements of the scattering coefficients. However, comparison of the relative intensities of the different polarizations in individual black-and-white and color composite images provides discriminatory mapping information. In the Death Valley, California, area, rough surfaces of young alluvial deposits produce strong responses at all polarizations. Smoother surfaces of older alluvial deposits show significantly lower responses. Evaporite deposits of different types and moisture contents have distinct polarization signatures. In the Wind River Basin, Wyoming, sedimentary rock units show polarization responses that relate to differences in weathering. Local intensity variations in like-polarization images result from topographic effects; strong cross-polarization responses denote the effects of vegetation cover and, in some cases, possible scattering from the subsurface. In the Savannah River Plant, South Carolina, forest cover characteristics are discriminated by polarization responses that reflect the density and structure of the canopy, and the presence or absence of standing water beneath the canopy.     

Potential Application of Multipolarization SAR for Pine-Plantation Biomass Estimation

This paper presents the techniques and the potential utility of multipolarization Synthetic Aperture Radar (SAR) data for pineplantation biomass estimation. Three channels of SAR data, one from the Shuttle Imaging Radar SIR-A and the other two from the aircraft SAR, were acquired over the Baldwin County, Alabama, study area. The SIR-A data were acquired with HH polarization and the aircraft SAR data with VV and VH polarizations. Linear regression techniques are used to estimate the pine-plantation biomass, tree height, and age using 21 test plots. The results indicate that the multipolarization data are highly related to the plantation biomass. The results suggest a potential application of multipolarization SAR for pine-plantation biomass estimation.     

SAR系统GMTI模式下目标特性分析和数据处理

根据SAR系统GMTI模式下点源目标(固定或运动)的特点,分析、计算了目标回波信号的多普勒质心、调频斜率和扫描波束的多普勒谱宽,讨论了载机平台位置和姿态的变化对系统的影响。文中按照分析和讨论的结论,提出了GMTI模式运动目标数据处理的解决方案。     

高分辨率多极化SAR图像目标分类误差分析

讨论了多极化SAR系统的极化误差及其对多极化SAR图像极化匹配目标分类性能的影响,并给出了计算结果。分析表明,多极化SAR系统的极化通道幅度不平衡误差对目标极化匹配结果的影响最大,交叉极化干扰项对此也有较大的影响。     

SAR图像特征数据立方体的构造

图像数据立方体的建立有助于图像数据的基于视觉内容的分析和多种知识的挖掘,包括汇总、分类、比较、关联和聚类.图像数据立方体对于图像分析是很有用的模型.实现一个维数很大的数据立方体是很困难的.但如果维数不够,则会使图像的建模范围过于粗糙、受到限制和不精确的扩展.因此,如何设计出既能满足效率要求,又能有足够的表达能力的图像数据立方体,是个亟待研究的问题.文中借鉴前人的思路和方法,构造了SAR图像立方体的模型.     

基于曲面拟合的SAR图像分维特征提取方法

本文对 ＳＡＲ图像的目标统计特性进行了分析,并对原始 ＳＡＲ图像数据采用小波分解进行了乘性噪声抑制,提出了一种用最小二乘法拟合空间曲面的分维特征提取算法,并应用该方法对ＳＡＲ图像数据进行了实验,得到了理想的效果．     

基于深度森林的多级特征融合SAR目标识别

大多数传统的合成孔径雷达(SAR)目标识别方法仅仅使用了单一的幅度特征,但是由于斑点噪声的存在,仅仅使用幅度特征会限制识别的性能。为了进一步提高SAR目标识别的性能,该文提出了一个基于深度森林的多级特征融合SAR目标识别方法。首先,在特征提取阶段,提取了多级幅度特征和多级密集尺度不变特征变换(Dense-SIFT)特征。幅度特征反映了目标反射强度,Dense-SIFT特征描述了目标的结构特征。而多级特征可以从局部到全局表征目标。随后,为了更完整、充分地反映SAR目标信息,借鉴深度森林的思想对多级幅度特征和多级Dense-SIFT特征进行联合利用。一方面通过堆叠的方式不断将多级幅度特征和多级Dense-SIFT特征进行融合,另一方面通过逐层的特征变换挖掘深层信息。最后利用得到的深层融合特征对目标进行识别任务。该文在MSTAR数据集上进行对比实验,实验结果表明所提算法在性能方面取得了提升,且其性能对超参数设置具有一定的鲁棒性。     

Forest Canopy Characterization and Vegetation Penetration Assessment with Space-Borne Radar

Synthetic Aperture Radar (SAR) images from the- National Aeronautics and Space Administration's Shuttle Imaging Radar-B (SIR-B) Mission were used to analyze the effects of radar incidence angle on information content and vegetation penetration. Three SAR data sets using incidence angles of 26°, 46°, and 58° were acquired over the mangrove jungles of Southern Bangladesh. The data sets were digitally processed using 3 × 3, 7 × 7, and 11 × 11 spatial filters and geometrically registered to a multisource-multilevel-corraborative data base consisting of Landsat data, forest map data, and in situ acquired forest enumeration and topographic information. Analyses revealed that significant vegetation "penetration" was found at all angles and that tree and canopy structural morphology may exert an influence on this phenomenon.     

双基线极化干涉合成孔径雷达的植被参数提取

由于单基线极化干涉SAR技术(如三阶段法、ESPRIT算法等)反演植被参数的方法受限于具有最小地体幅度比的观测数据,影响冠层或地表有效相位中心的估计,因此往往低估植被的高度。该文通过增加另一个基线的极化干涉SAR数据在一定程度上解决单基线反演技术存在的这个缺陷,并适当改进双基线极化干涉SAR技术,降低植被参数估计的动态偏差,最后用模拟数据和实际数据验证了该方法。     

基于深度森林的多级特征融合SAR目标识别

大多数传统的合成孔径雷达(SAR)目标识别方法仅仅使用了单一的幅度特征,但是由于斑点噪声的存在,仅仅使用幅度特征会限制识别的性能.为了进一步提高SAR目标识别的性能,该文提出了一个基于深度森林的多级特征融合SAR目标识别方法.首先,在特征提取阶段,提取了多级幅度特征和多级密集尺度不变特征变换(Dense-SIFT)特征.幅度特征反映了目标反射强度,Dense-SIFT特征描述了目标的结构特征.而多级特征可以从局部到全局表征目标.随后,为了更完整、充分地反映SAR目标信息,借鉴深度森林的思想对多级幅度特征和多级Dense-SIFT特征进行联合利用.一方面通过堆叠的方式不断将多级幅度特征和多级Dense-SIFT特征进行融合,另一方面通过逐层的特征变换挖掘深层信息.最后利用得到的深层融合特征对目标进行识别任务.该文在MSTAR数据集上进行对比实验,实验结果表明所提算法在性能方面取得了提升,且其性能对超参数设置具有一定的鲁棒性.     

一种三维森林场景极化SAR数据的快速模拟方法

该文提出了一种针对3维森林场景极化SAR数据模拟的快速模拟方法,解决了由于散射粒子数量巨大导致SAR回波模拟过程难以应用于大范围、高密度3维森林场景的问题.该方法通过将3维森林场景划分为“子场景”、“场景块”和“场景带”,构造出一系列虚拟散射粒子,在斜距误差较小的前提下将多个真实散射粒子对SAR回波信号的综合贡献用单个虚拟散射粒子等效,从而达到降低散射粒子数量,提高SAR回波信号模拟效率的目的.该文推导了3维森林场景切分各控制参数之间的被约束关系,分析了各参数的误差敏感度,得到了模拟精度和效率的优化组合准则.2.5×105 m2阔叶林场景的数据域相关性验证表明,该文提出的方法相比传统方法在保持较高模拟精度的前提下可以将模拟效率提高1个数量级.1.0×106 m2阔叶林场景的森林高度反演结果表明了该方法进行大范围森林场景SAR遥感数据模拟的有效性.     

面向目标特性精细提取的SAR数据融合成像处理方法

针对当前高分辨率星载SAR数据处理忽略目标相关散射特性随方位角变化而导致目标轮廓细节难以精细提取的问题,该文提出了一种面向目标特性精细提取的融合成像处理方法。首先,分析高分辨率条件下影响目标特性提取和SAR图像质量的因素,如轨道非线性、停-走模型、大气延时、高阶残留相位误差等,并提出针对性的补偿方法。在此基础上,建立基于目标时空谱特性的回波数学信号模型,解析目标散射特性时-空-频的变化规律,提出针对高分辨率、超大方位观测角范围的融合成像处理策略和方法,在距离-多普勒域、图像域通过融合处理提升SAR图像的品质。最后,通过对典型军事目标的仿真和成像处理,验证所提方法的有效性。      

Color- and Texture-Based Image Segmentation for Improved Forest Delineation

This paper concentrates on the delineation of forest boundaries from aerial images with focus on spatially contiguous and reproducible results for the Swiss National Forest Inventory. Because of the poor performance of common edge models to extract natural vegetation boundaries, this paper presents a combined method of image segmentation and wavelet-based texture features for the delineation of forest. The selected -measure-based segmentation method has been found to be useful to produce initial segmentation results, but lacks a semantic concept for forest vegetation. To overcome this conceptual limitation, the combination with wavelet transformation gives access to additional texture features and leads to a robust approach to obtain proper forest boundaries. Preliminary results are encouraging regarding the better agreement compared with maximum-likelihood classification results.     

Forest Biophysical Parameter Estimation Using L- and P-Band Polarimetric SAR Data

L- and P-band airborne polarimetric synthetic aperture radar (SAR) data acquired by the RAMSES system over different height maritime pine (Pinus Pinaster Ait.) stands of the Nezer forest (Landes, France) have been evaluated for forest biophysical parameter estimation. A pseudolinear correlation has been brought to evidence at P-band between polarimetric anisotropy and mean tree height, which is also linked to other biophysical parameters in the Nezer forest, meaning that SAR polarimetry constitutes a promising tool for forest parameter retrieval at low frequency. The spatial conditions have been evaluated through the quantification of the impact of signal-to-noise ratio diminution and resolution degradation on the forest height inversion. It has been shown that the inversion accuracy remains acceptable for N Esigma₀, representing the noise level of the SAR image, which is lower than -15 dB, and for spatial resolution increasing up to 15 m.     

基于简缩极化干涉SAR数据的森林垂直参数反演

森林区域竖直结构参数的反演是极化干涉雷达的一个重要应用,基于RVoG模型,运用全极化干涉数据可成功获得森林区域的地形估计及树高反演。该文将基于单基线简缩极化干涉SAR(C-PolInSAR)数据对森林区域进行林下地形估计及树高反演。推导单基线简缩极化干涉相干系数及相干区域,根据相干区域进行直线拟合,提出简缩极化干涉数据下的地形相位判别准则及体散射去相干估计方法,然后完成树高反演。通过L, P波段仿真数据以及实测机载数据对上述方法进行验证,获得正确地形及树高。简缩极化发射波极化状态不唯一,因此该文详细分析不同参数的椭圆极化对地形及树高等参数估计的影响,研究表明地形树高受椭圆极化波影响较小,也验证了估计方法的稳定性。     

SAR Data Filtering for Classification

This study compares several filtering techniqueprove to improve the classification of a synthetic aperture raar (SAR) image. With this aim in view, we separate thematic information and noise by speckle analysis, and we propose filters adapted to speckle models. The evaluation includes a brief theoretical analysis, a visual interpretation, a signal-to-noise ratio analysis, a comparison of required computer times, as well as classification results. The study also shows the effect of active microwaves on crop identification in the case where there is a one-date one-look SAR image, the utilization of X-and C-bands, and HH polarization.     

利用NOAA/AVHRR卫星资料进行海域表面特征提取

从７０年代开始，遥感技术获得了突飞猛进的发展，地球资源卫星获得的大量遥感图像中包含着有关地球表面，海洋表面等的大量信息，因此利用遥感图像卫星资料来研究地面及目标的光学特性，是一种极其有应用价值的方法和手段。ＮＯＡＡ／ＡＶＨＲＲ２有５个通道，通道１、和通道２位于可见光波段，用于观测可见光的反射太阳能，通道４和通道５是位于１０－１２μｍ的热红外通道，用于探测地面、海面及目标的辐射信息、通道３兼有上述两     

利用NOAA/AVHRR卫星资料进行海域表面特征提取

从70年代开始，遥感技术获得了突飞猛进的发展，地球资源卫星获得的大量遥感图像中包含着有关地球表面、海洋表面等的大量信息，因此利用遥感图像卫星资料来研究地面及目标的光学特性（热红外的反射率、辐射度及反射特性、温度及随时间的变化），是一种极其有应用价值的方法和手段。NOAA／AVHRR2有5个通道，通道1和通道2位于可见光波段，用于观测可见光的反射太阳能，通道4和通道5是位于10～12μm的热红外通道，用于探测地面、海面及目标的辐射信息。通道3兼有上述两种特性，是一个比较复杂的通道。我们仅选用了通道4和通道5的遥感卫星图像进行了研究，从黑体辐射公式出发，反演出传感器的探测温度Td，并对其进行了传感器定标校正、温度的大气校正，从而得到比较精确的表面温度Ts。同时利用双红外通道4和5相关度大的特性，对其进行了组合，用回归统计分析法进行了反演精度比较。研究结果表明，本文的技术和方法可以进行大面积区域目标特征的提取，把传统的点研究推广到面的研究。     

面向植被边坡的地基SAR高相干点选择

基于高相干点进行相位分析,才能有效保证地基SAR（Synthetic Aperture Radar,合成孔径雷达）形变测量的准确性。在差分干涉测量领域中,广泛使用幅度离差法,可以有效地选择出岩石、建筑物等PS（Permanent Scatterer,永久散射体）点作为高相干点,但对于植被边坡,采用该方法选择出高相干点数量较少,不利于差分干涉处理。在星载SAR领域,普遍采用StaMPS方法解决植被边坡的高相干点选择问题。本文探索了StaMPS方法在地基SAR领域的适用性,提出采用非PS点来计算相干系数门限,并引入DS（Distributed Scatterer,分布式散射体）选择技术,进一步提高了高相干点的数量。对一处植被边坡的实验结果表明,相比于幅度离差法和StaMPS方法,改进方法在提高高相干点数量的同时,有效保证了其相位质量。      

Model-Associated Forest Parameter Retrieval Using VHF SAR Data at the Individual Tree Level

A simple statistical extension at the individual tree level for an earlier-developed very high frequency forest backscatter model is proposed. This extended model treats trunk volumes as random quantities. A concept of random forest reflection coefficient is also introduced to characterize radar returns from individual trees. Based on the extended model, a set of algorithms for estimating the mean trunk (stem) volume from synthetic aperture radar data at the individual tree level is developed assuming that the areal tree density is known. The algorithms are specified for different scenarios related to a priori information on parameters of statistical distributions for the trunk volume and fluctuations of the forest reflection coefficient. An approximate lower bound on the standard deviation in the unbiased estimation of the mean trunk (stem) volume is proposed. This bound can be readily obtained by means of computer simulation for any specified statistical distribution for the trunk volume and fluctuations of the forest reflection coefficient. Performance analysis for the proposed algorithms is numerically performed by means of Monte Carlo simulation for a variety of scenarios. This analysis has shown that the algorithms provide nearly unbiased and efficient estimates, and the proposed lower bound is a very accurate approximation. The results of the study have demonstrated that the approach and methods developed in this paper suggest promising solutions in accurate forest parameter retrieval.