一种动态分区的边境狭长区域网平差方法

针对内陆边境狭长区域特殊自然环境、气候条件和其他敏感因素引起的区域网平差精度问题,提出一种利用动态分区规划的区域网平差方法。基于高分立体像对影像,利用层次分析法结合不同尺度DEM实现任务区动态分区,将控制信息以最优方式传递至无控区域,并逐步剔除粗差解算出最优区域网平差模型。实验结果表明,该方法生成的边境狭长区DEM高程均方根误差小于7.5m,DOM产品单点定位精度误差小于5m,丘陵与平原区域误差小于1m,相邻影像之间的几何拼接精度优于0.5m。     

资源三号卫星影像DEM提取与精度分析

为提高国产遥感卫星影像定位精度,提取高精度DEM,该文基于卫星遥感的成像模型和间接平差理论,利用有理函数模型和像面仿射变换模型,构建了卫星影像平差数学模型;利用该数学模型对资源三号卫星立体影像进行了平差处理,建立了研究区域DEM。分析结果表明:DEM在无控制点情况下,平地区域平面精度达到6.12m,高程精度达到8.39m;山地区域平面精度达到6.20m的,高程精度达到8.62m,数据精度能够满足1∶50000比例尺地形图测图标准和要求。研究结果验证了国产卫星数据的定位精度。     

基于机载LiDAR的高寒山区遥感高程数据精度评估

遥感高程数据是获取缺资料地区DEM（Digital elevation models）数据的重要手段。然而,由于高寒山区实地高程测量稀少,难以对多源遥感DEM数据进行统一验证。ICESat-2等新的遥感高程数据在高寒山区也缺乏相应的精度评估。针对此问题,以青藏高原东北缘的冰沟流域作为研究区,采用机载航空遥感获取的大范围LiDAR（Light Detection And Ranging）DEM数据对新产品ICESat-2 ATL06（Ice, Cloud, and Land Elevation Satellite-2, Land Ice Height）、ALOS DEM（12.5 m分辨率）以及新版本SRTM V3（SRTM Arc-Second Global 1 V003）、ASTER GDEM V3（ASTER Global DEM）进行验证,并分析地形因子与均方根误差RMSE的关系。研究结果表明：ICESat-2 ATL06数据在高寒山区的RMSE为0.747 m。由于其较高的精度,可用于验证缺资料地区的其他遥感高程数据。其他遥感高程数据的精度都相对较低,ALOS 12.5 m数据的RMSE为5.284 m;ASTER GDEM V3版本的RMSE为9.903 m。实验所采用的4种遥感高程数据与机载LiDAR DEM均具有较高的相关性,相关系数在0.998与1.000之间。实验还揭示了坡度是影响遥感DEM精度的主要因素。除ICESat-2 ATL06外,其他高程数据的RMSE均随坡度的增大先减小再增大,且都存在一个最佳坡度值。鉴于地形复杂多样的冰沟流域具有青藏高原高寒山区的典型特征,多源遥感DEM数据在该区域的验证结论具有较好的代表性,可为相似地区DEM数据的使用和评估提供重要的知识补充。     

星载SAR干涉技术获取DEM及其精度分析

星载合成孔径雷达干涉（InSAR）技术是一种数据覆盖范围广、廉价、高效、方便的数字高程模型（DEM）获取方法,但在地面植被覆盖广、大气水汽含量高的地区其影像相干性随时间基线的增加迅速降低;同时,SAR卫星的轨道误差也影响DEM精度。利用ERS-1/2卫星串行模式SAR数据获取镇江地区DEM,分析了轨道误差对DEM精度的影响;根据干涉相位的统计特性,从理论上给出干涉相位噪声与相干系数和视数之间的关系。实验结果表明就干涉像对的卫星轨道误差和相位噪声而言,在小区域内DEM精度优于3.5 m。     

多重观测卫星影像的无控区域网平差

为了保证卫星遥感影像获取地面目标区域信息的准确性,需要对影像数据进行几何校正,几何校正的精度决定着遥感图像应用的效果。常规的校正方法需要地面控制点信息,但在境外、我国西部、荒漠等这些地方很难获取控制点。针对上述问题,提出了多重观测卫星影像的无控区域网平差方法,提高无控定位的精度。与常规区域网平差不同,该方法的误差方程基于"单物方—多像方"的连接点集建立,在无控条件下误差方程能够收敛到更加精确的解。通过误差补偿模型调整每幅影像的有理函数模型系数,使得不同方向的定位误差进行抵消,从而在无控制点的条件下提高遥感影像的定位精度。首先,利用卫星影像的RPC文件和像方误差补偿模型,建立多重重叠的区域网平差模型,然后,利用共轭梯度算法迭代求解误差方程,最后调整RPC系数,提高遥感图像定位精度。通过资源三号卫星影像试验表明:多重观测区域网平差将遥感图像的平面定位精度由19.8m提高至12.9m,能够有效提高影像的几何定位精度。     

基于已有控制资料的正射影像自动更新

高分辨率遥感影像作为地理国情监测的基础数据,其新数据获取后的变化检测与适时更新是地理国情实现动态监测的关键。该文将已有的正射影像和数字高程模型作为新获取高分辨率卫星影像的控制资料,结合通用成像模型的有理多项式模型,首先实现高分辨率卫星影像与已有DOM/DEM之间的特征提取和匹配,然后利用RFM模型的区域网平差对匹配结果进行粗差剔除,从而实现高分辨率卫星影像的对地定位,最终实现正射影像更新。实验证明该方法可提高匹配精度与定向参数的可靠性。     

应用Kriging插值方法插补干涉雷达DEM奇异值斑块

干涉合成孔径雷达（InSAR)可以获取地表的三维信息，但是由于雷达数据的热噪声等原因，使得干涉雷达获取的DEM存在高程奇异值斑块，严重影响了InSAR获取DEM的应用。鉴于高程奇异值斑块的DEM中分布的随机性，以及一般DEM数据具有的空间连续性和空间自相关性，针对这一问题，笔提出了利用基于空间统计学的Kriging插值法插补DEM的高程奇异值斑块。本对STRM提供的DEM进行了插补，同时对插补的结果进行了精度分析。结果表明可以满足DEM的应用精度要求。     

非量测相机倾斜航空影像空三加密精度评价

鉴于倾斜空三加密及其精度评价工作的重要性,使用非量测型相机获取倾斜航空影像,利用计算机视觉的SIFT特征匹配算子对影像的空三加密进行改进,然后对加密结果进行精度评价。结果表明:对于人工构造物密集地区的立体影像,特征点的数量密集,空三加密精度倾斜影像检查点的精度相较下视影像精度高;影像的分辨率为10cm,量测结果平面x、y的中误差都在精度范围中,高程z值中误差略大;增加控制点数目后,平面x与y的精度明显提高。分析影响试验结果的误差来源有:航摄质量、控制点布设与量测精度、内业人员操作、区域网平差方法等。     

InSAR技术获取高山峡谷区DEM精度研究

通过对金沙江河段高山峡谷区L波段的Alos-Palasar和C波段的Radarsat-2雷达单视复数数据的干涉处理,获取此区域的数字高程模型(DEM)。利用SRTM 90m分辨率的DEM为参考数据,通过对比分析发现InSAR技术生成的DEM精度与相干系数、地形和波长有密切的关系。同时也验证了在相干性好,地形起伏不太剧烈的地区,用InSAR技术生成DEM是可行的。     

低山丘陵区多源数字高程模型误差分析北大核心CSCD

作为多学科交叉与渗透产物的数字高程模型(DEM)已在诸多学科和领域及实际应用中发挥了重要作用,但目前能够免费获取的高分辨全球DEM在不同区域仍存在很大的不确定性,应用之前进行质量评估至关重要。以烟台市为实验区,以大比例尺地形图(1∶10 000)生成的DEM为参照,结合坡度、坡向和土地覆被类型等地学因子,定量分析了目前广泛应用的两个版本ASTER GDEM(先进星载热辐射和反射辐射计全球数字高程模型)ASTETR 1和ASTER 2及不同空间分辨率SRTM DEM(航天飞机雷达地形测绘任务)(SRTM 1:～30m和SRTM 3:～90m)在低山丘陵区高程、坡度及坡向误差。结果表明:在研究区域内,ASTER 1、ASTER 2、SRTM 3、SRTM 1总体高程均方根误差分别为8.7m、6.3m、3.7m和2.9m。ASTER与SRTM的高程精度不同程度地受坡度、坡向以及土地覆被类型等地学因子的影响,DEM误差随坡度增加而增大,其中SRTM 3精度对该因子最敏感。尽管坡向对DEM精度影响不明显(4种DEM在不同坡向上的均方根误差波动范围均不超过2m),但是不同土地覆被类型下这4种DEM精度差异显著。此外,分析4种DEM提取的坡度可知,SRTM 1的均方根坡度误差最低(2.5°)、ASTER 1与ASTER 2的坡度的均方根误差大致相同(3.6°、3.9°)、SRTM 3的坡度均方根误差最高(4.3°)。坡向的精度SRTM 1最高,ASTER 1与ASTER 2次之,SRTM 3最低。研究结果对我国低山丘陵区ASTER GDEM与SRTM DEM的应用与精度评估具有一定的借鉴作用。     

Joint interferometric calibration based on block adjustment for an airborne dual-antenna InSAR system

Mapping large areas using airborne dual-antenna interferometric synthetic aperture radar (InSAR) usually requires processing and mosaicking of different scenes from multiple strips. The overlapping areas of these multiple strips should have consistent elevation values. Due to the unstable attitude of the plane, the interferometric parameters usually vary for each scene during mapping. Therefore, interferometric calibration technology for high-precision height retrieval is required for the correction of the interferometric errors. The traditional interferometric calibration methods for a single scene usually use ground control points (GCPs) to estimate the interferometric parameters – this method cannot guarantee a consistent height in the area of overlap. Besides, GCPs are difficult to deploy over rough terrain, making it impossible to use traditional calibration methods. In this article, a joint interferometric calibration method based on the block adjustment theory used in photogrammetry is proposed for airborne dual-antenna InSAR. This method considers the accurate digital elevation model (DEM) height reconstruction model and can be applied with sparse GCPs. The principle of the proposed method is to make the best use of the GCPs within all the scenes and the tie points (TPs) between the adjacent scenes to establish an error relationship model. First, the weighting values of all GCPs and TPs based on their retrieval elevation error caused by the interferometric phase error and the position distribution difference are introduced in the proposed method. Next, the interferometric parameters are weighted to reduce the condition number of the normal equation. Then, an alternative approximation approach combined with the sparse matrix decomposition technique LDL^T is utilized to solve the normal equation, and the corrected interferometric parameters for each scene are obtained. High-precision joint interferometric calibration results for airborne InSAR systems are achieved by the proposed method and validated by experiment. Using the proposed method, the average mean error (AME) and root mean square error (RMSE) are below 0.6037 and 0.9176 m, respectively. Meanwhile, the maximum AME and RMSE of the reconstructed DEM height difference for the validation TPs in the overlapped area of the adjacent scenes are reduced from 1.2909 and 1.7245 m to 0.8864 and 1.2087 m, respectively.     

Multi-Observed Block Adjustment for Satellite ImagesWithout Ground Control Points

In order to ensure the accuracy of target area’s ground information,it is necessary to do geometric correction for remote sensing images.Geometric correction has a great influence on the application of remote sensing images.Traditional geometric correction needs ground control points.However,it is difficult to obtain ground control points in some places,such as abroad,western China and desert.To improve positioning accuracy without ground control points,multi\|overlapping block adjustment model is built.Different from traditional method,error equations are built depending on multi\|projection in image space of a single object.In this way,error equations can converge to more accurate solutions.By adjusting the rational polynomial coefficients of each image,the positioning errors in different directions are compensated to a certain extent.Thus the positioning accuracy of remote sensing images is improved.First,block adjustment model and error compensation model are built with RPC coefficients.Then,conjugate gradient algorithm is used to solve the error equations iteratively.Finally the RPC coefficients are adjusted to improve the accuracy of positioning without ground control points.The ZY\|3 data test shows that multi\|overlapping block adjustment model increase the plane positioning accuracy of remote sensing images from 19.8m to 12.9m and the method can effectively improve the absolute positioning accuracy of remote sensing images.     

DEM generation and error analysis using the first Chinese airborne dual-antenna interferometric SAR data

Terrain survey with traditional photogrammetry is often difficult in western China, such as Qingzang tableland at an average height of 5000 m above sea level and the southwest China area with cloudy weather. To resolve western terrain mapping, the first Chinese single-pass airborne Interferometric Synthetic Aperture Radar (InSAR) system was successfully developed by the Institute of Electronics, Chinese Academy of Sciences (IECAS) in 2004. The main objective of this article is to examine and evaluate the performance of the airborne SAR system through interferometric processing and error analysis. First, the article describes how high-precision digital elevation models (DEMs) are derived from the airborne dual-antenna (single-pass) InSAR data. In order to improve the precision, the antenna eccentricity correction and parameter calibration with the least square method (LSM) are proposed. Based on the airborne dual-antenna InSAR bore-sight model, this article summarizes the primary factors that influence the accuracy of DEMs in data processing, and analyses the errors induced by these factors. Then, the global positioning system (GPS)/inertial measurement unit (IMU) data, acquired and stored by the position and orientation system (POS), is used for analysing the quantitative relationships among the platform height, baseline length, baseline angle, look angle and DEM error. The experimental data used are airborne dual-antenna X-band InSAR data, and the measured ground control points (GCPs) are used to validate the accuracy of the DEM. The evaluation results in terms of the standard deviation (SD) and the average mean error (AME) are derived by comparing the reconstructed InSAR DEM with the reference GCPs. The AMEs of the X-direction, the Y-direction and the height are up to 2.078, 9.149 and 1.763 m, respectively. The SDs of the X-direction, the Y-direction and the height are up to ±1.379, ±0.764 and ±1.086 m, respectively. These results agree with the previously calculated quantitative errors. The error value of the Y-direction seems too large, a possible result of system errors. In general, the airborne dual-antenna InSAR system initially meets the requirements of 1:50 000 terrain mapping in western China.     

Cross-validation as a means of investigating DEM interpolation error

Studies of the detailed characteristics of DEM error have been hampered by the difficulty in obtaining a large sample of error values for a DEM. The approach proposed in this paper is to resample a DEM to a lower resolution and then reinterpolate back to the original resolution which produces a large sample of error values well distributed across the DEM. This method is applied to a sample area from Scotland, which contains a variety of terrain types. The results show that the standard measure of error, the root mean square error (RMSE) of elevation, shows only moderate correlation with a visual assessment of the quality of DEMs produced by a range of interpolation methods. The frequency distribution and strength of spatial autocorrelation are shown to vary with the initial data density and interpolation method. When the source data density is low, the error has strong spatial autocorrelation and a distribution that is close to being Gaussian. However, as the data density increases, levels of spatial autocorrelation drop and the distribution becomes leptokurtic with values very strongly clustered around zero. At the level of the individual DEM point, elevation error is shown to be a poor predictor of error in slope derivatives which depend on the spatial pattern of elevation errors around the point and are also sensitive to differences in terrain. At the level of a whole DEM, however, RMSE of elevation is a good predictor of RMSE in gradient and aspect but not of curvature.     

SRTM DEM enhancement using a single set of PolSAR data based on the polarimetry-clinometry model

ABSTRACT

The freely available global and near-global digital elevation models (DEMs) have shown great potential for various remote sensing applications. The Shuttle Radar Topography Mission (SRTM) data sets provide the near-global DEM of the Earth’s surface obtained using the interferometry synthetic aperture radar (InSAR). Although free accessibility and generality are the advantages of these data sets, many applications require more detailed and accurate DEMs. In this paper, we proposed a modified and advanced polarimetry-clinometry algorithm for improving SRTM topography model which requires only one set of polarimetric synthetic aperture radar (PolSAR) data. The azimuth and range slope components estimation based on polarization orientation angle (POA) shifts and the intensity-based Lambertian model formed the bases of the proposed method. This method initially compensated for the polarimetry topography effect corresponding to SRTM using the DEM-derived POA. In the second step, using a modified algorithm, POA was obtained from the compensated PolSAR data. The POA shifts by the azimuth and range slopes’ variations based on the polarimetric model. In addition to the polarimetric model, a clinometry model based on the Lambertian scattering model related to the terrain slope was employed. Next, two unknown parameters, i.e. azimuth and range slope values, were estimated in a system of equations by two models from the compensated PolSAR data. Azimuth and range slopes of SRTM were enhanced by PolSAR-derived slopes. Finally, a weighted least-square grid adjustment (WLSG) method was proposed to integrate the enhanced slopes’ map and estimate enhanced heights. The National Aeronautics and Space Administration Jet Propulsion Laboratory (NASA JPL) AIRSAR was utilized to illustrate the potential of the proposed method in SRTM enhancement. Also, the InSAR DEM was employed for evaluation experiments. Results showed that the accuracy of SRTM DEM is improved up to 2.91 m in comparison with InSAR DEM.     

一种机载InSAR水体阴影的提取和识别方法

为了解决机载InSAR DEM中水体和阴影区域质量不佳需要区分修复的问题,提出一种综合利用机载InSAR数据源自动提取水体和阴影并加以识别的方法。首先基于InSAR DEM进行粗差点检测,利用粗差点作为种子点在SAR图像中区域生长,提取完整的水体和阴影区域;然后利用沿斜距向高程差和雷达俯角构造约束条件自动识别两者。通过对实测的机载高分辨率InSAR数据进行处理,水体阴影的识别率达到92%以上,其中水体和地形阴影的识别较好,而受制于DEM内在噪声等因素的影响,由树木造成的小块阴影容易造成误分。     

基于M型标准块的视觉引导机器人系统手眼标定方法研究

针对线结构光传感器引导的机器人系统的手眼标定问题,提出了一种以M型标准块为标定物的方法。该M型标定物的两条平行的脊线作为约束,基于两条平行脊线的约束建立包含手眼关系、机器人运动学以及两条直线位姿参数误差的模型。首先基于定点约束求解手眼关系初值并以此为基础解算出直线位姿参数的初值,然后通过最小二乘法解算误差参数并补偿到模型中,不断迭代直至计算的误差参数小于阈值,最终得到最终的机器人手眼关系及运动学误差参数。为了验证标定方法的有效性,以某精加工平面为被测物,利用线结构光机器人系统对平面进行测量,得到平面点云;拟合最小二乘平面,计算点到平面距离的均方根值作为评价依据。分别对所述M型标准块和标准球两种方法进行了实验对比,结果表明,相较于标准球方法,所述M型标准块方法得到的均方根误差由0.152 mm减少到0.080 mm,均方根误差的标准差由0.043 mm减少到0.005 mm,其标定结果的精度及稳定性得到显著提高。     

Planar block adjustment and orthorectification of Chinese spaceborne SAR YG-5 imagery based on RPC

Yaogan-5(YG-5), launched in December 2008, is a Chinese high-resolution spaceborne synthetic aperture radar (SAR) satellite, with a ground resolution of 3 m. However, the direct geometric positioning accuracy of YG-5 slant range images is low and so is the mosaic accuracy of the orthoimages. To improve the geometric accuracy of YG-5 orthoimages, this article proposes a strategy to calculate the rational polynomial coefficients for each SAR image and then uses a planar block adjustment method to solve for the orientation research parameters of the SAR images to achieve the orthorectification while a auxiliary digital elevation model is necessary for height constraint. Compared with the traditional orthorectification method using a single image, this strategy can ensure both uniformity in positioning accuracy of orthorectified images and high mosaic accuracy of adjacent orthoimages based on a small number of ground control points (GCPs). Tests using Chinese YG-5 satellite data over Xi’an and Xianning, China show that, using four GCPs positioned in the four corners of the test area, we can achieve independent check point plane accuracies better than ±4 m after the planar block adjustment. Finally, this article demonstrated that seamless mosaic geometry levels can be attained after the block orthorectification.     

基于UWB的LSM-Taylor级联车辆定位算法

为了满足智能车在室内的高精度定位要求,针对室内的伪三维定位场景,提出了一种基于超宽带(Ultra Wideband,UWB)的LSM-Taylor级联车辆定位算法.该算法以到达时间差(Time Difference of Ar-rival,TDOA)为定位方式,以多基站最小二乘法(Least Square Method,LSM)定位算法的计算结果为初始值,通过Taylor级数迭代估计车辆的精确位置.该算法主要解决多径效应和非视距产生的测量误差对定位精度的影响,从而提高定位精度.在仿真结果中,相比LSM定位算法,LSM-Taylor级联定位算法的定位结果分布更加紧密,定位精度更高.实际测试结果表明,该定位算法的均方根误差(Root Mean Squared Error,RMSE)在10 cm以下,能满足智能驾驶中的室内定位要求,验证了该方法的有效性.