Accuracy of differential shift estimation by correlation and split-bandwidth interferometry for wideband and delta-k SAR systems

Estimation of differential shift of image elements between two synthetic aperture radar (SAR) images is the basis for many applications, like digital elevation model generation or ground motion mapping. The shift measurement can be done nonambiguously on the macro scale at an accuracy depending on the range resolution of the system or on the micro scale by employing interferometric methods. The latter suffers from phase cycle ambiguities and requires phase unwrapping. Modern wideband high-resolution SAR systems boast resolutions as small as a few tens of a wavelength. If sufficiently many samples are used for macro-scale shift estimation, the accuracy can be increased to a small fraction of a resolution cell and even in the order of a wavelength. Then, accurate absolute ranging becomes precise enough to support phase unwrapping or even make it obsolete. This letter establishes a few fundamental equations on the accuracy bounds of shift estimation accuracy for several algorithms: coherent speckle correlation, incoherent speckle correlation, split-band interferometry, a multifrequency approach, and correlation of point scatterers in clutter. It is shown that the performance of split-band interferometry is close to the Crame/spl acute/r-Rao bound for a broad variety of bandwidth ratios. Based on these findings, Delta-k systems are proposed to best take advantage of the available radar bandwidth.     

一种基于曲线SAR的三维目标特征提取与自聚焦新算法

曲线合成孔径雷达（Curvilinear Synthetic Aperture Radar,简称曲线SAR）是一种新的三维成像模式，通过让载机作曲线飞行，曲线SAR系统能同时在方位维和高度维形成大的合成孔径。本文提出一种曲线SAR的三维目标特征提取和自聚焦新方法。新方法是建立在一种灵活的接收数据模型之上的参数化的新方法。仿真实验证明了所提方法的有效性。     

多基线机载合成孔径雷达影像匹配的SANCC法

为了有效利用合成孔径雷达(synthetic aperture radar,SAR)多基线影像的几何信息和辐射信息,提高匹配精度,提出了一种适用于多基线SAR幅度影像的自适应归一化互相关系数和(sum of adaptive normalized cross-correlation,SANCC)影像匹配方法。该方法首先利用SAR成像参数、平台参数和物方高程范围构建匹配方向线;然后引入Gestalt原理的接近性和相似性原则对窗口内的像素加权,计算获得沿匹配方向线的SANCC值;最后采用winner-take-all(WTA)优化策略获取多基线影像的匹配结果和物方三维信息。利用国产机载SAR系统获取的多基线影像进行匹配试验,结果表明与常规的相关匹配方法相比,该方法可以获得更为密集、精确的匹配点,有效减少了由重复纹理造成的误匹配,并能较好地解决由纹理匮乏导致的匹配难题。     

Map Drift算法在实时成像处理器中的应用

随着计算机技术的发展，实时成像处理器技术的成熟，不但处理数据量越来越大，而且对成像质量要求也越来越高，实时成像处理中需要有自动聚焦算法。但是由于实时成像处理器的特殊结构和实时性要求，需要算法的计算量不能太大、不能太不确定。因此并不是所有的算法都适合用于实时处理。 主要介绍Map Drift算法的原理和实现步骤以及在实时成像处理器中的应用。文中对适合于实时成像处理器中应用的公式进行了推导，得到频域二次相位误差系数和相关峰位置的线性关系，并利用该关系在频域进行相位补偿，有效缩减迭代次数。利用一种插值来计算相关峰位置的方法，消除了离散化计算相关峰位置所带来的量化误差的影响，并在迭代过程中改进二次相位误差系数和相关峰位置关系的系数，对减少迭代次数，提高估计精度有一定帮助。最后给出仿真结果，验证算法的正确性。     

基于最小二乘平差的全极化SAR配准偏移量估计方法

全极化合成孔径雷达影像(PolSAR)相对单极化SAR影像有更加丰富的地表信息。为了提高SAR影像偏移量跟踪技术估算偏移量的精度,提出一种基于最小二乘平差法的全极化SAR数据偏移量估计方法。首先利用全极化SAR不同极化通道数据估算偏移量得到多个观测值,然后通过最小二乘平差法对多余观测值循环剔除粗差得到最优偏移量。文中选取美国科罗拉多州湖城(Lake city)区域的两景JPL无人机UAVSAR全极化SAR影像进行实验,结果表明,文中新方法具有良好的粗差探测和去除功能,相对于已有研究结果,在方位向和距离向的偏移量估算精度都有明显提高,达到15%~25%。新方法提高了偏移量跟踪估算偏移量的精度,这对利用偏移量跟踪技术监测滑坡、地震及冰川等有重要的意义。     

SAR稀疏成像模型参数自适应选择

合成孔径雷达(SAR)稀疏成像模型中的参数选择对于SAR稀疏成像的性能有重要影响,也是当前SAR稀疏成像研究中的难点问题。已有参数选择方法普遍存在适用于个别模型或者运算量大的缺点。基于最大后验概率估计和贝叶斯推理,提出了一种无需额外先验信息的自适应参数选择方法,所有需要的参数都可从已知的数据中获取。通过推导得到模型参数与信号、噪声方差的关系,避免了对数据进行一系列的训练处理,因此极大地减小了计算量。仿真数据和实测数据处理表明,本文方法在实现了较为精确的参数优化选择的前提下,其计算量远低于贝叶斯信息论准则、L-曲线等已有参数选择方法。     

区域Gamma混合模型的SAR图像分割

针对传统Gamma混合模型用于SAR图像分割时忽略像素间空间相关性,导致分割结果不连续并产生大量误分割的现象,提出了区域Gamma混合模型的SAR图像分割算法。首先对图像进行分水岭分割,得到过分割区域块,然后将其作为输入样本进行基于Gamma混合模型的聚类,在模型的参数估计过程中进一步考虑区域间的空间相关性,设计邻域因子融入到迭代过程,得到邻域加权类分布概率。该算法充分利用像素间的空间相关性,能够降低噪声对分割结果的影响。通过合成图像和真实SAR图像的实验表明,本文算法能够实现SAR图像的准确分割。     

极化SAR图像中非平稳目标的检测与区分方法研究

在合成孔径雷达(SAR)极化测量中,通常认为雷达视角相对于目标是不变的。然而实际上SAR通常具有较宽的方位向波束,在孔径合成过程中,所成的全分辨率SAR图像包含多个斜视角下的目标回波。极化参数的子孔径分析可以用来对成像过程表现出非平稳散射行为的媒质进行检测。本文提出了一套基于子孔径分解和对散射矩阵及参数进行统计分析的方法,来检测两类非平稳目标。仿真结果表明,这种方法具有较好的非平稳目标检测和区分性能。     

Modeling and sensitivity analysis of navigation parameter errors for airborne synthetic aperture radar stereo geolocation

For the high-resolution airborne synthetic aperture radar (SAR) stereo geolocation application, the final geolocation accuracy is influenced by various error parameter sources. In this paper, an airborne SAR stereo geolocation parameter error model, involving the parameter errors derived from the navigation system on the flight platform, has been put forward. Moreover, a kind of near-direct method for modeling and sensitivity analysis of navigation parameter errors is also given. This method directly uses the ground reference to calculate the covariance matrix relationship between the parameter errors and the eventual geolocation errors for ground target points. In addition, utilizing true flight track parameters’ errors, this paper gave a verification of the method and a corresponding sensitivity analysis for airborne SAR stereo geolocation model and proved its efficiency.     

Topographic correction of ALOS-PALSAR images using InSAR-derived DEM

Topographic corrections of synthetic aperture radar (SAR) images over hilly regions are vital for retrieval of correct backscatter values associated with natural targets. The coarse resolution external digital elevation models (DEM) available for topographic corrections of high resolution SAR images often result into degradation of spatial resolution or improper estimation of backscatter values in SAR images. Also, many a times the external DEMs do not spatially co-register well with the SAR data. The present study showcases the methodology and results of topographic correction of ALOS-PALSAR image using high resolution DEM generated from the same data. High resolution DEMs of Jaipur region, India were generated using multiple pair SAR images acquired from ALOS-PALSAR using interferometric (InSAR) techniques. The DEMs were validated using differential global positioning system measured elevation values as ground control points and were compared with photogrammetric DEM (advanced spaceborne thermal emission and reflection radiometer – ASTER) and SRTM (Shuttle Radar Topography Mission) DEM. It was observed that ALOS-PALSAR images with optimum baseline parameters produced high resolution DEM with better height accuracy. Finally, the validated DEM was used for topographic correction of ALOS-PALSAR images of the same region and were found to produce better result as compared with ASTER and SRTM-DEM.     

几何失真对SAR图像匹配的影响及校正方法分析

合成孔径雷达SAR用于图像匹配导航系统,可大大提高导航精度。然而地形起伏和大气干扰等因素使得SAR实时图存在几何失真,降低了匹配性能。通过对SAR实时图几何失真成因分析,提出了一种二维可分离的几何校正方法,并得出方位向几何失真可跟随成像处理预先校正,图像后处理只需校正距离向。利用数字高程模型DEM和正射图模拟得到了3幅具有不同地形特征的SAR实时图,并分别利用3种几何校正方法对其进行统计实验,得到了不同地形SAR实时图几何校正方法选取的准则。     

Maximum-likelihood estimation for multi-aspect multi-baseline SAR interferometry of urban areas

The reconstruction of digital surface models (DSMs) of urban areas from interferometric synthetic aperture radar (SAR) data is a challenging task. In particular the SAR inherent layover and shadowing effects need to be coped with by sophisticated processing strategies. In this paper, a maximum-likelihood estimation procedure for the reconstruction of DSMs from multi-aspect multi-baseline InSAR imagery is proposed. In this framework, redundant as well as contradicting observations are exploited in a statistically optimal way. The presented method, which is especially suited for single-pass SAR interferometers, is examined using test data consisting of experimental airborne millimeterwave SAR imagery. The achievable accuracy is evaluated by comparison to LiDAR-derived reference data. It is shown that the proposed estimation procedure performs better than a comparable non-statistical reconstruction method.     

顾及分类器参数的全极化SAR图像特征选择与分类

全极化SAR获取的信息量远多于传统SAR,但信息量的增加并不能确保分类精度的提高,如何有效进行特征选择至关重要。针对自适应特征选择问题,提出一种顾及分类器参数的特征选择和分类方法。该方法以支持向量数为评估依据,结合遗传算法进行特征选择,并同时对分类器参数进行寻优;最后利用优选的特征集和模型参数进行分类。为验证算法的有效性,利用两组全极化数据进行了监督分类实验。实验结果表明,提出方法降低了SVM分类器对自身参数的敏感性,而且能在较少特征个数下具备良好的泛化性能,分类精度优于未经过特征选择和参数优化的方法。     

混合智能优化算法的SAR图像特征选择

为提高SAR图像自动目标识别的准确率及实时性,提出了一种基于混合智能优化的SAR图像特征选择算法。首先,采用分形特征对SAR图像进行增强,基于分割后的图像提出了一种基于图像矩的方位角估计方法。然后基于未校正和校正后的图像分别提取Zernike矩、Gabor小波系数和灰度共生矩阵构成候选特征集合,使用遗传算法结合二值粒子群的混合优化算法实现SAR图像特征选择。最后,采用MSTAR数据库验证本文算法的有效性。实验结果表明,优化后的特征集合具有一定泛化能力,一方面提高了SAR目标识别的准确率,另一方面减小了SAR图像目标识别的时间。     

An unsupervised segmentation method based on MPM for SAR images

An unsupervised segmentation method for synthetic aperture radar (SAR) images is proposed. It alternately approximates the maximization of the posterior marginals estimate of the pixel class labels and estimates all model parameters except the number of classes during segmentation. In this method, a multilevel logistic (MLL) model for the pixel class labels and Gamma distribution for the marginal distribution of each class in the observed SAR image are employed. In our implementation, the expectation-maximization algorithm is used to estimate parameters of the Gamma distributions, and the iterative conditional estimation algorithm is used to estimate the MLL model parameters. The segmentation results for synthetic and real SAR images show that the proposed method has a good performance.     

Delta-K interferometric SAR technique for snow water equivalent (SWE) retrieval

This letter describes the concept of using delta-K technique on interferometric synthetic aperture radar (InSAR) data for deriving the snow water equivalent (SWE) of dry snow-covered ground by utilizing the presence of scatterers in both datasets. The main scattering contribution from a dry snow cover is from the snow-ground interface. Thus, the interferometric phase of two SAR images, one with no snow and one with dry snow cover, contains information on the SWE. By performing a delta-K processing of the two SAR scenes followed by averaging, an estimation of the SWE can be achieved. The first step in the delta-K InSAR processing is to split the band into two nonoverlapping subfrequency band images. The resulting two subband images then contain two new carrier frequencies with a small delta frequency or delta-K separation. The next step is to multiply the two subband images together to obtain the delta-K image, one for summer and one for winter. Finally, the delta-K interferometric SAR image is generated by multiplying the two delta-K images from summer and winter together. In this letter, experimental results using European Remote sensing Satellite 1 (ERS-1) data from a summer and winter situation show that the delta-K phase can be estimated to a few degrees accuracy for an area of 10/spl times/10 km/sup 2/ corresponding to an SWE accuracy of approximately 100 mm.     

利用SVM的全极化、双极化与单极化SAR图像分类性能的比较

支持向量机(SVM)以其在小训练样本时良好的分类性能,目前已广泛应用于多个领域.本文在极化SAR图像特征提取基础上,将SVM应用于极化SAR图像分类,定性和定量地比较了全极化、双极化和单极化SAR图像的分类性能,分析了不同的极化组合对分类结果的影响,并根据地物极化散射特性分析了分类精度差异的成因.实测极化SAR数据的实验结果表明,全极化数据能获得最好的分类性能,双极化次之,单极化最低,且在某些情况下,双极化与全极化分类性能接近.     

相干变化指数法在震害快速评估中的应用

合成孔径雷达(SAR)具有全天候全天时的特点,为快速展开地震救援,评估震害损失提供了重要的数据支持。本文基于汶川震中东南平原地区的ENVISATASAR影像,利用相位干涉方法处理得到相干变化指数图像,对6个评估区域的像元相干变化指数进行统计分析,以相干变化指数平均值作为区分建筑物基本完好与损毁的阈值,并与实地调查结果相比,精度基本达到70%。验证干涉相干变化指数法对震害快速评估有效的同时,也为震区提供震害救援的依据。     

Estimation of Radial Velocity of Moving Targets by Along-Track Interferometric SAR Systems

Along-track interferometric synthetic aperture radar (AT-InSAR) can be used to estimate the radial velocity of ground moving targets, starting from interferometric phase measures. The estimation obtained from a single-phase interferogram suffers from ambiguities. To solve these problems, multichannel AT-InSAR systems are required. In this letter, we analyze the radial velocity maximum-likelihood estimation accuracy with respect to AT-InSAR system parameters, such as velocity values and different clutter and thermal noise levels. We consider two different models for the target response: a deterministic model and a zero-mean Gaussian model. The presented results show that AT-InSAR systems exhibit better estimation accuracies for low-velocity values (slow targets).