Stereo analysis of high-resolution SAR images for building height estimation in cases of orthogonal aspect directions

SAR stereo image analysis for 3D information extraction is mostly carried out based on imagery taken under same-side or opposite-side viewing conditions. For urban scenes in practice stereo is up to now usually restricted to the first configuration, because increasing image dissimilarity connected with rising illumination direction differences leads to a lack of suitable features for matching, especially in the case of low or medium resolution data. However, due to two developments SAR stereo from arbitrary viewing conditions becomes an interesting option for urban information extraction. The first one is the availability of airborne sensor systems, which are capable of more flexible data acquisition in comparison to satellite sensors. This flexibility enables multi-aspect analysis of objects in built-up areas for various kinds of purpose, such as building recognition, road network extraction, or traffic monitoring. The second development is the significant improvement of the geometric resolution providing a high level of detail especially of roof features, which can be observed from a wide span of viewpoints. In this paper, high-resolution SAR images of an urban scene are analyzed in order to infer buildings and their height from the different layover effects in views taken from orthogonal aspect angles. High level object matching is proposed that relies on symbolic data, representing suitable features of urban objects. Here, a knowledge-based approach is applied, which is realized by a production system that codes a set of suitable principles of perceptual grouping in its production rules. The images are analyzed separately for the presence of certain object groups and their characteristics frequently appearing on buildings, such as salient rows of point targets, rectangular structures or symmetries. The stereo analysis is then accomplished by means of productions that combine and match these 2D image objects and infer their height by 3D clustering. The approach is tested using real SAR data of an urban scene.     

基于图像特征的星载SAR图像模拟研究

SAR图像模拟技术被广泛应用于SAR系统的设计和验证、SAR图像的正射纠正、雷达图像解译和目标识别等。随着星载SAR的发展,必然面临着对星载SAR图像模拟的大量需求。本文首先从SAR图像的几何特征和辐射特征出发,探讨了SAR图像模拟技术的原理,分析了RD（Rang Doppler）模型,后向散射模型和斑噪模型。在传统RD模型的基础上,根据不同地形特征（起伏地形和平坦地形）考虑不同的后向散射模型。特别强调了在平坦地形情况下,需要地物分类数据的参与,并利用Ulaby和Dobson的后向散射模型。另外,在SAR图像统计特征的基础上,进行SAR图像的乘性噪声模拟,可以满足更逼真的SAR场景需求。然后,给出了图像模拟的算法流程,并对关键步骤的算法做了分析。最后,在实现基于图像特征的星载SAR图像场景模拟算法的基础上,选择新疆窝依牙地区和天津地区分别进行起伏地形和平坦地形的模拟试验,实验结果证明了本文模拟算法的有效性。     

地形三维建模中特殊地物的处理方法

在三维建模领域,空中三角测量通过少量的控制点,按一定的数学模型加密解算出所需要的全部控制点及每张像片的外方位元素并由此进行三维重建工作;若遇到一些特殊地物(如桥梁、铁塔、水域等)时,往往会出现模型错乱、模糊不清、有漏洞或无法正确显示等一系列问题。本文采用倾斜摄影测量建模的方式,对一些特殊地物进行信息采集,再利用专业软件进行加工处理,用3dsmax类建模软件对采集的信息进行精细建模,用skyline类软件进行三维场景编辑及漫游,从而能够实现基于原本样貌重建特殊地物的三维模型。     

使用ERS-1/2干涉测量SAR数据生成DEM

干涉合成孔径雷达（ＩＮＳＡＲ）数据已被证明能生产精确的数据高程模型（ＤＥＭ），我们已开发了从单视ＳＡＲ复影像数据自动生成数字高程模型的新软件，基于ＳＡＲ多视强度影像的最小二乘曩像匹配被用于复影像对的配准，达到很高的配准精度（０．０１～０．０５像元精度）。一种新的计算目标点３维坐标（Ｘ，Ｙ，Ｚ）的方程还被提出，卫星轨道，姿态和基线参数以及相位常数被纳入在方程中并被表示了时间的线性函数，利用至少６个地面控制点能够同时估算这些参数，本文还给出了意大利埃特地区ＥＲＳ－１／２ＳＡＲ数据处理结果。     

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。     

<Emphasis Type="Italic">3D</Emphasis> Modelling of Urban Area Using Synthetic Aperture Radar (SAR)

Synthetic aperture radar (SAR) is a newly-developed remote sensing technology that works in all weather and independent of daylight. Recent satellite designs such as TerraSAR-x, which have resolutions of a couple of meters and sub-meters, have provided appropriate data for modelling and monitoring of urban areas. Image classification and height information extraction is possible considering the nature of SAR data. In this paper, a proper classification method for high-resolution SAR images has been used in urban areas. This classifier is based on statistical models. First, statistical models that are well adapted to urban SAR images are selected. Initial labelling is performed using the maximum likelihood method. A method based on Markov random fields is applied to improve the results by considering neighbourhood information. Meanwhile, topographic information is extracted using the phase difference obtained from SAR interferometry. After classification and height extraction, the homogeneous regions consisting of locations with similar objects are determined. The homogeneous region adjacency graph are generated using vectors containing classification information, extracted objects, height of pixels forming each region, and information on the neighbouring areas. Height and classification information are then merged by assigning height conditions based on the nature of objects and optimizing an energy function. The results obtained, including buildings, streets, and corner reflectors, are easily recognizable. The overall accuracy is improved from 57% in the initial classification to 95% in the employed procedure. Moreover, the accuracy of height estimation is about 2.74 m, which is acceptable for height estimations of buildings with more than one floor.     

附带星历参数的星载合成孔径雷达影像自动配准算法

Fourier-Mellin不变性对称相位匹配滤波器,即symmetric phase-only matched flitering of fourier-mellin invariant(SPOMF-FMI)能解决影像配准的平移、旋转和放缩几何变换问题;然而,该方法在实际应用中具有局限性,包括图像几何尺度和图像非线性几何畸变处理能力.为了克服上述缺陷,提出了一种附带星载参数的星载合成孔径雷达(synthetic aperture radar,SAR)影像自动配准算法.该方法采用由粗到精匹配策略,由距离-多普勒影像粗匹配和改进SPOMF-FMI影像精匹配组成.使用Radarsat-1和ENVISAT ASAR影像做实验,结果表明本方法在处理重复轨道或相同升、降轨星载SAR影像(即确保相似影像纹理测度)配准问题时,能达到子像素级的配准精度.     

基于Kalman滤波的InSAR基线估计方法

针对目前SAR干涉测量中基线估计现存的问题,提出了利用Kalman滤波和配准参数进行基线估计的方法.所提出的方法具有不需地面控制点、不受地形限制和不依赖于轨道参数等优点,并可以估计时变的基线参数.利用南京地区的ERS-1/2 tandem数据进行了试验研究,并对提出的方法进行了验证.结果表明,在精确的卫星轨道数据和地面控制点不能获取时,所提出的方法仍能有效地估计InSAR基线.这在一定程度上补偿了轨道偏移带来的误差,为获取高精度的DEM奠定了基础.     

用于末制导的SAR图像多子区实时匹配算法

为实现用于无人飞行器末制导的SAR实时影像与多源基准影像的匹配,提出一种新的基于子区位置关系约束的SAR图像匹配算法。首先将SAR图像进行对数变换,并进行小波包分解,在高频子图部分对小波包系数进行收缩,抑制噪声,然后利用断面检测法提取边缘。匹配中将SAR实时图分为多个子区,用子区灰度标准差和频谱直方图判别其可匹配性,并更新不可匹配区。将各子区位置分布关系用作匹配结果的约束条件。为保证算法实时性,匹配采用基于提升小波的金字塔分层搜索策略。用星载SAR图像和卫星光学图像进行了大样本匹配实验,实验结果表明该方法是一种能在精度、实时性和可靠性方面都能满足导弹末制导要求的实时匹配算法。     

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。     

基于有理函数模型的星载SAR影像几何校正

主要研究了面向星载合成孔径雷达SAR(Synthetic Aperture Radar)影像几何处理的有理函数模型RFM(Rational Function Model)的建模、求解和应用方法。首先采用与地形无关的方式,利用严密的距离-多普勒模型(Range Doppler,RD)构建虚拟控制点格网来进行RFM建模,实现了一种RFM模型参数的快速无偏解算新方法,能够取得相对于RD模型很高的拟合精度,在几何定位功能上实现了对RD模型的有效替代,同时大大提高了计算效率。然后在此基础上利用RFM模型实现了星载SAR影像的快速几何校正,为了提高几何校正结果的绝对定位精度,引入少量地面控制点对RFM模型进行了像方改正处理,以消除SAR影像几何定位的系统误差,并利用ENVISAT ASAR数据的实验结果验证了本文方法的有效性。     

Indoor scene reconstruction using feature sensitive primitive extraction and graph-cut

We present a method for automatic reconstruction of permanent structures, such as walls, floors and ceilings, given a raw point cloud of an indoor scene. The main idea behind our approach is a graph-cut formulation to solve an inside/outside labeling of a space partitioning. We first partition the space in order to align the reconstructed models with permanent structures. The horizontal structures are located through analysis of the vertical point distribution, while vertical wall structures are detected through feature preserving multi-scale line fitting, followed by clustering in a Hough transform space. The final surface is extracted through a graph-cut formulation that trades faithfulness to measurement data for geometric complexity. A series of experiments show watertight surface meshes reconstructed from point clouds measured on multi-level buildings.     

利用主题模型的遥感图像场景分类

提出了一种基于主题模型与特征组合相结合的遥感图像分类方法。该方法首先对图像进行尺度不变特征变换(SIFT)、几何模糊特征(GB)和颜色直方图特征(CH)提取,接着利用潜在概率语义分析(pLSA)模型分别对所得到的图像特征进行潜在主题的挖掘,然后对所得到的主题概率特征进行组合,最后利用支持向量机(SVM)分类器进行场景分类。实验表明,与传统分类方法相比,主题模型更具优势;与使用单特征相比,特征组合具有更高的分类准确率。     

SARBuD 1.0： 面向深度学习的GF-3 精细模式SAR建筑数据集

合成孔径雷达SAR（Synthetic Aperture Radar）是开展城市建筑区信息获取与动态监测的重要数据源。本文建立了一个面向深度学习建筑区提取的中高分辨率SAR建筑区数据集SARBuD1.0 （SAR BUilding Dataset）。该数据集包含了覆盖中国不同区域的27景高分三号（GF-3）精细模式SAR图像,并从中获取了建筑区共计60000个SAR样本数据,结合光学图像与专家解译,制作了与样本数据对应的标签图像。SARBuD1.0数据集包含了不同地形场景类型、不同分布类型、不同区域的建筑区。该数据集可支持研究者对建筑区进行图像特征分析、辅助图像理解,并可对当前热点深度学习方法提供训练、测试数据支持。本文以山区建筑为例,使用传统纹理特征与深度学习特征对建筑区进行了特征分析与比较,相比于传统的人工设计的纹理特征,卷积神经网络具有更深、更多的特征,利用网络模型浅层的不同卷积核采样可得到各种纹理特征,在网络的深层卷积结构中可获取代表着类别的深层语义特征,使得分类器能更好地检测并提取图像中指定的目标。基于本数据集利用深度学习方法对不同地形区域的建筑区进行提取实验。实验结果表明基于本数据集训练的深度学习模型,对建筑区提取可以取得良好的结果,说明该数据集可以很好支持面向大数据的深度学习方法。其他学者可以基于SARBuD1.0数据集开展建筑区图像特征分析与语义分割提取等方面的研究。     

高分三号卫星全极化SAR影像九寨沟地震滑坡普查

基于光学遥感影像的区域滑坡普查易受云雾天气的影响,存在滑坡体调查不全面的问题,无法满足震后应急调查与恢复重建的需求。本文提出了一种极化SAR卫星数据滑坡普查方法,采用高分三号全极化SAR卫星影像数据,以九寨沟地震震区为实验区,在深入分析滑坡体和其他地物类型散射特征的基础上,融合极化特征、纹理特征和地形特征等多维特征信息,结合高分二号影像获取的训练样本,构建基于BP神经网络的全极化SAR数据滑坡自动识别模型,实现滑坡体的自动快速识别。与高分辨率光学影像与无人机航空影像目视解译结果相比较,总体识别精度为92.8%,Kappa系数为0.715,识别准确度满足地震应急实际应用的需求。研究成果可用于震区大区域滑坡体的普查,为后续开展无人机高分辨率影像滑坡体详查、灾后应急与景区恢复提供辅助信息支撑,并促进国产高分SAR卫星数据在防震减灾中的应用。     

基于少量控制点的Radarsat-2影像快速几何纠正技术研究

Radarsat-2卫星依据其搭载的GPS接收机可实现3倍中误差小于60m的精确实时定轨。由此本文提出依据其影像元数据信息实现快速几何纠正的方法,该方法利用少量的几个控制点来消除Radarsat-2影像与待纠正参考系间的系统误差。从而实现Radarsat-2影像的快速几何纠正。本文并依据SAR斜距成像原理的纠正公式和实地采集的GPS数据,验证了元数据中所提供RPF模型的内部精度和外部符合精度。通过实验验证了本快速纠正技术可以获得中误差小于2个像素的平面几何纠正精度。     

稀少控制的多平台星载SAR联合几何定标方法

几何定标采用地面控制点获取距离-多普勒模型中的精确几何参数,用于完成星载SAR影像高精度几何定位。但在广域范围内,特别是高山地区域,控制点极难获取。此外,传统定标方法仅面向单一平台SAR影像,尚不能实现多平台影像的联合几何定标。针对上述问题,本文提出一种基于稀少控制的多平台星载SAR联合几何定标方法。该方法从包含实测控制点的主影像出发,使用点位追踪算法获取主影像与从影像之间的连接点,并以连接点为桥梁逐级完成从影像的几何定标。本文采用京津冀地区南北向分布共计235 km的3景TerraSAR-X、3景TanDEM-X、5景高分三号影像进行联合几何定标试验,仅使用5个控制点即完成了所有影像的几何定标,并利用SF-3050星站差分GNSS接收机采集实测GPS点进行精度评价。结果表明使用稀少控制点定标后的TSX/TDX影像的几何定位精度优于3 m,GF-3影像的几何定位精度优于7.5 m,验证了该方法的有效性和正确性。     

Classification of dual- and single polarized SAR images by incorporating visual features

Fully and partially polarimetric SAR data in combination with textural features have been used extensively for terrain classification. However, there is another type of visual feature that has so far been neglected from polarimetric SAR classification: Color. It is a common practice to visualize polarimetric SAR data by color coding methods and thus it is possible to extract powerful color features from such pseudo color images so as to gather additional crucial information for an improved terrain classification. In this paper, we investigate the application of several individual visual features over different pseudo color generated images along with the traditional SAR and texture features for a novel supervised classification application of dual- and single-polarized SAR data. We then draw the focus on evaluating the effects of the applied pseudo coloring methods on the classification performance. An extensive set of experiments show that individual visual features or their combination with traditional SAR features introduce a new level of discrimination and provide noteworthy improvement of classification accuracies within the application of land use and land cover classification for dual- and single-pol image data.