Integrated Use of Optical and Radar Satellite Data for Active Faults and Corresponding Displaced Landforms in Kozani Basin (Greece)

Abstract

The usefulness of integration of SAR (ERS-1) and Landsat TM data for study active faults and the corresponding displaced landforms in flat or almost flat areas has been demonstrated. The study area is the Kozani basin which in May 13, 1995 was affected by a strong earthquake (Ms=6.6). After co-registration and resampling the two data sets were merged to form a combined image. The combined image offers the spectral characteristics of the TM data with the spatial resolution and roughness sensitivity of SAR images. The merging method used was the IMS to RGB transform. The criteria and parameters examined were geomorphic features, drainage network analysis, slope processes, terrain analysis, and observations on spatial distribution of soil cover as well as linear features that correspond to fracture zones crossing the basin. The use of the combined image allowed us to identify tectonic terraces in the basin produced by activity of normal faults located in the adjacent relief zone.     

机载合成孔径雷达技术在地形测绘中的应用及其进展

合成孔径雷达(SAR)技术由于其全天候、全天时的工作能力,成为近几年来摄影测量与遥感领域的研究热点。机载SAR系统的逐渐增多为SAR技术的发展提供了丰富的数据源。本文综述了近两年来国内外机载合成孔径雷达(AIRSAR)技术在地形测绘中的研究、试验和应用进展,评述了机载合成孔径雷达关键技术的解决途径,展望了机载合成孔径雷达技术在地形测绘和其他领域中的应用前景。     

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.     

面向对象高可信SAR数据精确处理

针对合成孔径雷达（synthetic aperture rdar,SAR）数据在地形地物、森林植被等方面的处理与解译难题,介绍了近年来利用多角度、多波段、多极化、极化干涉等多模态航空航天SAR数据,建立基于散射机理的地物特性知识库,构建地形辐射校正模型、极化干涉处理模型、立体测量模型、基于知识的解译模型等,开发出高分辨率机载极化干涉SAR数据获取系统和SAR影像高性能解译软件系统,实现了精度高、可靠性强、识别类型丰富的SAR影像高可信处理与解译的原理、技术与方法,同时对成果在测绘、林业等行业的应用情况进行了介绍,对研究中存在的问题和解决思路进行了探讨。     

A method for vertical adjustment of digital aerial photogrammetry data by using a high-quality digital terrain model

The accuracy of vertical position information can be degraded by various sources of error in digital aerial photogrammetry (DAP) based point clouds. To address this issue, we propose a relatively straightforward method for automated correction of such point clouds. This method can be used in conjunction with any 3D reconstruction method in which a point cloud is generated from a pair of aerial images. The crux of the method involves separately co-registering each DAP point cloud (formed by the overlap of two or more images) to a common airborne laser scanning (ALS) based digital terrain model. The proposed method has the following essential steps: (1) Ground surface patches are identified in the normalized DAP point clouds by selecting areas in which standard deviation of vertical height is low, (2) height differences between the DAP and ALS point clouds are calculated at these patches, and (3) a correction surface is interpolated from these height differences and is then used to rectify the entire DAP point cloud. The performance of the proposed method is verified using plot data (n = 250) from a forested study area in Eastern Finland. We observed that DAP data from the area corrected using our proposed method resulted in significant increases in prediction accuracy of key forest variables. Specifically, the root mean squared error (RMSE) values for dominant height predictions decreased by up to 23.2%, while the associated model R² values increased by 16.9%. As for stem volume, RMSEs dropped by 20.6%, while the model R² improved by 14.6%, respectively. Hence, prediction accuracies were almost as good as with ALS data. The results suggest that vertically misaligned DAP data, if rectified using an algorithm such as the one presented here, could deliver near ALS data quality at a fraction of the cost.     

Determination of complexity factor and its relationship with accuracy of representation for DEM terrain

Based on the estimating rule of the normal vector angles between two adjacent terrain units, we use the concept of terrain complexity factor to quantify the terrain complexity of DEM, and then the formula of terrain complexity factor in Raster DEM and TIN DEM is deduced theoretically. In order to make clear how the terrain complexity factor E _CF and the average elevation h affect the accuracy of DEM terrain representation RMSE _Et , the formula of Gauss synthetical surface is applied to simulate several real terrain surfaces, each of which has different terrain complexity. Through the statistical analysis of linear regression in simulation data, the linear equation between accuracy of DEM terrain representation RMSE _Et , terrain complexity factor E _CF and the average elevation h is achieved. A new method is provided to estimate the accuracy of DEM terrain representation RMSE _Et with a certain terrain complexity and it gives convincing theoretical evidence for DEM production and the corresponding error research in the future.     

A Neural Network Landuse Classifier for SAR Images Using Textural and Fractal Information

Abstract

Artificial neural networks (ANN) have recently been popularly used in image classification. Input features to most ANNs are extracted based on a one class per pixel basis. This requires a large number of training samples and thus a slow training rate. In this paper, we describe the use of a windowing technique to extract textural features such as average intensity, second moment of intensity histogram and fractal surface dimension from an image. This method of image characterization reduces the number of training samples efficiently, yet retains a reasonable overall classification accuracy. The ANN is trained based on the back‐error propagation algorithm. The method is applied for landuse classification of Synthetic Aperture Radar (SAR) images. An example is given for a site in Kedah State, Malaysia. The SAR images (HH,HV,VV) were taken by the Canadian Centre for Remote Sensing (CCRS) CV‐580 airborne C‐band SAR system in November 1993 during their GlobeSAR mission in Malaysia. These multi‐polarization SAR images are co‐registered with a Landsat Thematic Mapper (TM) channel 5 image from same area. An overall classification accuracy of about 86.95% is achieved using windowing technique, as compared to 68.22% based on one class per pixel approach. This shows that through fractal and textural information, the windowing technique when applied in an ANN classifier has a great potential in remote sensing applications.     

Unambiguous Reconstruction and High-Resolution Imaging for Multiple-Channel SAR and Airborne Experiment Results

Azimuth ambiguity occurs in synthetic aperture radar (SAR) systems due to the well-known constraint of minimum antenna area, particularly at high resolutions and wide swaths. A space time domain method can be utilized to remove this ambiguity if the multiple-channel data are available. In this letter, a modified approach is presented to determine the filter weight vectors. This approach was successfully applied to the real data, which were collected by an experimental airborne multiple-channel SAR system. The channel imbalance and the error in antenna phase center position are analyzed in detail.     

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。     

Investigation for planimetric accuracy of airborne synthetic aperture radar

The airborne SAR images were tested for geometric accuracy in order to assess the suitability of present airborne radar systems for topographic mapping. Images were transformed to terrain coordinate system using 2-D conformal, affine and polynomial transformations. Standard error in positional discrepancies at check points show that the geometric fidelity of present airborne SAR system is compatible with planimetric mapping requirements at 1∶50,000 and smaller scale.     

高分辨率机载SAR影像判读实验

近年来机载SAR系统在国内得到快速发展。由于SAR系统具有不同于光学传感器的成像机理,因而可以提供全新的数据获取手段和不同的地物信息。利用高分辨率机载SAR影像进行解译时必须熟悉其成像机理和图像信息特点,充分了解SAR图像解译标志的特点和各类地物的解译规律。本文以高分辨率机载SAR数据为基础,对比光学影像,从SAR成像特点和地物散射特性出发,对SAR影像进行了目视解译,分析了SAR影像在地形地物要素识别中的应用潜力。     

机载双天线InSAR复图像自动配准研究

干涉雷达复图像之间的精确配准是提高雷达干涉测量精度的关键之一。对于机载双天线干涉雷达系统,精确的地形高程测量需要精确的干涉相位来保证,从而要求干涉复图像在亚像元级的精确配准。为了使两幅机载双天线InSAR复图像之间配准精度达到亚像元级以及提高运算效率,在分析前人算法的基础上,分别采用基于快速傅里叶变换(FFT)复相关精配准算法以及基于过采样图像的精配准方法对机载双天线InSAR复图像进行精配准试验和比较,生成了干涉条纹图,并进行了相干性分析。试验结果表明,机载双天线复图像相干性很高、偏移量较小;对于机载双天线复数据而言,这两种方法都可以有效地实现复图像数据的高精度配准,但是从配准精度以及算法运行时间来看,基于FFT的复相关精配准算法较优,运算效率更高,且更具实用性。     

融合升降轨的极化干涉SAR三层模型植被高度反演方法

森林参数的获取不仅可以估算地表生物量和林下地形,还有助于研究全球碳循环和分析全球气候变化。极化干涉SAR植被参数反演算法一般是基于随机地体两层模型(RVoG),但是当实际植被有着冠层、树干层和地表层的明显三层结构时,植被参数反演精度就会变差;另外,由于机载SAR系统数据的近距远距垂直向波数差异较大,导致试验结果存在着由其引起的系统误差。针对这两个问题,本文提出了一种融合升降轨的极化干涉SAR三层模型植被参数反演方法。该方法首先采用三层植被RVoG模型修正微波在穿透植被时的散射过程;然后采用融合升降轨道数据的方式削弱其系统误差;最后,采用非线性迭代平差的反演算法来进行植被高度反演。为了验证该方法的有效性,采用了德国宇航局DLR提供的BioSAR2008项目的两景升轨及两景降轨E-SAR P波段全极化SAR数据进行试验,并采用3组反演策略进行比较分析。结果表明,三层植被模型能够更好地描述植被散射过程;同时,新方法有效降低了由垂直向波数引起的系统误差,提高了树高反演精度。     

Fine-resolution forest tree height estimation across the Sierra Nevada through the integration of spaceborne LiDAR,airborne LiDAR,and optical imagery

ABSTRACT

Forests of the Sierra Nevada (SN) mountain range are valuable natural heritages for the region and the country, and tree height is an important forest structure parameter for understanding the SN forest ecosystem. There is still a need in the accurate estimation of wall-to-wall SN tree height distribution at fine spatial resolution. In this study, we presented a method to map wall-to-wall forest tree height (defined as Lorey’s height) across the SN at 70-m resolution by fusing multi-source datasets, including over 1600 in situ tree height measurements and over 1600?km² airborne light detection and ranging (LiDAR) data. Accurate tree height estimates within these airborne LiDAR boundaries were first computed based on in situ measurements, and then these airborne LiDAR-derived tree heights were used as reference data to estimate tree heights at Geoscience Laser Altimeter System (GLAS) footprints. Finally, the random forest algorithm was used to model the SN tree height from these GLAS tree heights, optical imagery, topographic data, and climate data. The results show that our fine-resolution SN tree height product has a good correspondence with field measurements. The coefficient of determination between them is 0.60, and the root-mean-squared error is 5.45?m.     

Preliminary results of the establishment of a marine geodetic point in the Pacific Ocean

In November 1968, a marine geodetic control point was established in the Pacific Ocean at a water depth of6,200 feet. The control point (reference point) consists of three underwater acoustic transponders, two of which are powered with lead-acid batteries and the third with an underwater radioisotope power source “URIPS” with a10- to20- year life expectancy. Four independent measuring techniques (LORAC airborne line-crossing, satellite, ship inertial, and acoustic techniques) were used to measure and determine the coordinates of the control point. Preliminary analysis of the acoustic and airborne data indicates that high accuracies can be achieved in the establishment of geodetic reference points at sea. Geodetic adjustment by the method of variation of coordinates yielded a standard point error of±50 to±66 feet in determining the unknown ship station. The original location of the ship station as determined by shipboard navigation equipment was off by about1,600 feet. Paper previously published in the Proceedings of the Second Marine Geodesy Symposium of the Marine Technology Society.     

Evaluation and perspectives of using multitemporal L-band SAR data to monitor deforestation in the Brazilian Amazo/spl circ/nia

Japanese Earth Resources Satellite 1 (JERS-1) synthetic aperture radar (SAR) data were evaluated to map areas of deforestation in a Brazilian Amazo/spl circ/nia test-site. The results were compared with information derived from a Landsat TM multitemporal series. Unambiguous detection of deforested areas was observed only when the entire deforestation process (slash, burning, and terrain clearing) had already occurred. This result recommends further investigations on the effectiveness of horizontal polarization SAR data to map deforestation in a consistent basis. The cross-polarized (horizontal-vertical) channel designed to be in the ALOS/PALSAR system is expected to improve the distinction between forested and recently deforested areas.     

SAR interferometric coherence analysis for snow cover mapping in the western Himalayan region

Abstract

Information of snow cover (SC) over Himalayan regions is very important for regional climatological and hydrological studies. Precise monitoring of SC in the Himalayan region is essential for water supply to hydropower stations, irrigation requirements, and flood forecasting. Microwave remote sensing has all weather, day and night earth observation capability unlike optical remote sensing. In this study, spaceborne synthetic aperture radar interferometric (InSAR) coherence analysis is used to monitor SC over Himalayan rugged terrain. The feasibility of monitoring SC using synthetic aperture radar (SAR) interferometry depends on the ability to maintain coherence over InSAR pair acquisition time interval. ERS-1/2 InSAR coherence and ENVISAT ASAR InSAR coherence images are analyzed for SC mapping. Data sets of winter and of snow free months of the Himalayan region are taken for interferogram generation. Coherence images of the available data sets show maximum decorrelation in most of the area which indicates massive snowfall in the region in the winter season and melting in the summer. Area showing coherence loss due to decorrelation is mapped as a snow-covered area. The result is validated with field observations of snow depth and it is found that standing snow is inversely related to coherence in the Himalayan region.     

Earth observations during space shuttle flight STS‐41: Discovery's mission to planet earth

Abstract

Multi‐temporal ERS‐1 SAR data acquired over a large agricultural region in West Bengal was used to classify kharif crops like rice, jute and sugarcane. Rice crop grown under lowland management practice showed a temporal characteristic. The dynamic range of backscatter was highest for this crop in temporal SAR data. This was used to classify rice using temporal SAR data. Such temporal character was not observed for the other study crops, which may be due to the difference in cultivation practice and crop calendar. Significant increase in backscatter from the ploughed fields was used to derive information on onset and duration of land preparations. Synergistic use of optical remote sensing data and SAR data increased the separability of rice crop from homesteads and permanent vegetation classes.     

Gravity field terrain effect computations by FFT

The widespread availability of detailed gridded topographic and bathymetric data for many areas of the earth has resulted in a need for efficient terrain effect computation techniques, especially for applications in gravity field modelling. Compared to conventional integration techniques, Fourier transform methods provide extremely efficient computations due to the speed of the Fast Fourier Transform (FFT. The Fourier techniques rely on linearization and series expansions of the basically unlinear terrain effect integrals, typically involving transformation of the heights/depths and their squares. TheFFT methods will especially be suited for terrain reduction of land gravity data and satellite altimetry geoid data. In the paper the basic formulas will be outlined, and special emphasis will be put on the practial implementation, where a special coarse/detailed grid pair formulation must be used in order to minimize the unavoidable edge effects ofFFT, and the special properties ofFFT are utilized to limit the actual number of data transformations needed. Actual results are presented for gravity and geoid terrain effects in test areas of the USA, Greenland and the North Atlantic. The results are evaluated against a conventional integration program: thus, e.g., in an area of East Greenland (with terrain corrections up to10 mgal), the accuracy ofFFT-computed terrain corrections in actual gravity stations showed anr.m.s. error of0.25 mgal, using height data from a detailed photogrammetric digital terrain model. Similarly, isostatic ocean geoid effects in the Faeroe Islands region were found to be computed withr.m.s. errors around0.03 m