利用特征组合检测算法的无人机遥感影像匹配研究

影像匹配是无人机遥感影像拼接和三维建模的基础和关键步骤。结合不同算法的优势,本文提出一种基于特征组合与RANSAC算法的无人机遥感影像匹配方法。该匹配方法首先采用AKAZE算法检测影像的特征点,然后利用SIFT描述符描述特征向量并获取特征点的主方向,最后基于单映射变换矩阵的RANSAC算法进行精准匹配。本文对基于特征组合与RANSAC算法的匹配效果进行了试验对比分析,试验结果表明：与常用匹配方法的匹配效果相比,本文的匹配方法继承了AKAZE算法的快速匹配能力,匹配总耗时介于AKAZE算法和SIFT算法之间,约为BRISK算法匹配耗时的20%;同时,该匹配方法继承了SIFT算法的多匹配点对性能,从整体匹配效果来看,本文的匹配方法优于AKAZE、SIFT、BRISK算法。     

基于点云空间索引的隧道断面提取方法

探索了一种基于三维激光点云空间索引的隧道断面提取方法。该方法主要通过建立空间索引,利用k维树（kdimensional tree,KD树）和规则格网提取频数分布特征,采用含有多种几何模型约束条件的随机抽样一致（random sample consensus,RANSAC）算法完成断面拟合。实验结果表明,该方法对隧道点云的平均降噪率达到97%,能准确提取隧道断面。     

基于Harris与RANSAC算法的无人机影像拼接方法研究

针对Harris算法在进行无人机影像拼接时的特征点误匹配问题,利用RANSAC算法对误匹配的特征点进行剔除,进而实现无人机影像的无缝拼接。首先,基于Harris算法提取兴趣点,利用最短欧式距离进行特征匹配;然后,利用RANSAC算法实现对特征点的精确匹配;最后,利用RANSAC算法得到的单应性矩阵完成无人机影像拼接。实验结果表明:本文方法能够较好地剔除无人机影像拼接时误匹配的特征点,实现对无人机影像的拼接,拼接效果良好。     

基于主方向的SIFT误匹配点剔除方法

针对RANSAC方法在剔除SIFT误匹配点方面的不足,提出了一种基于主方向的SIFT误匹配点剔除方法。结果表明,该算法与RANSAC相比,耗时较少,剔除效果更好,保留的正确匹配点数更多,且当误匹配点数量多于正确匹配点时也能得到很好的剔除效果。     

基于主方向的 SIFT 误匹配点剔除方法简

针对RANSAC 方法在剔除SIFT 误匹配点方面的不足,提出了-种基于主方向的SIFT 误匹配点剔除方法.结果表明,该算法与RANSAC 相比,耗时较少,剔除效果更好,保留的正确匹配点数更多,且当误匹配点数量多于正确匹配点时也能得到很好的剔除效果.     

一种三维激光点云中建筑物立面渐进分割方法

针对建筑物立面分割的问题,该文提出了一种三维激光点云中建筑物立面渐进分割方法。建立三维格网索引,通过分析建筑物立面在三维格网内的空间分布特征和二维平面格网内投影的线性分布特征,确定立面种子格网和投影线的拟合点,拟合投影线并基于种子格网约束生长完成每层格网中立面粗分割;使用RANSAC算法对粗分割后的立面点云进行面拟合,实现精细分割,并将各层格网中的立面进行合并,实现建筑物立面的完整分割。实验结果表明,该方法能有效地实现建筑物立面的精细化分割,有助于后续的建筑物精细三维模型重建。     

一种喀斯特地貌无人机遥感影像的快速匹配方法

针对喀斯特地貌地形起伏大、沟壑纵深、影像局部存在阴影,该地区无人机遥感影像匹配难度大,该文提出一种基于AKAZE特征和RANSAC算法的喀斯特地貌无人机遥感影像的快速匹配方法。该方法采用AKAZE特征检测算法提取喀斯特地貌无人机遥感影像的特征点,能够较为精简地描述出影像的特征,影像匹配时间缩短,匹配速度提高明显;由于预匹配存在错误点对,利用单应性矩阵的RANSAC算法进行精准匹配,剔除错误匹配点对,同时得到最优的单应性矩阵。实验结果表明,该方法的匹配耗时和匹配正确率较优于经典的SURF和BRISK算法,该匹配方法更适合于喀斯特地貌无人机遥感影像匹配。     

机载激光雷达数据的建筑物三维模型重建

机载激光雷达技术可以快速获取建筑物高精度高密度的三维空间信息.文章提出一种建筑物三维模型重建方法,该方法利用栅格数据使用RANSAC算法提取建筑边界上的长线段,采用规则化策略获取建筑轮廓;在模型生成阶段,基于原始点云和屋顶拓扑关系约束,获取屋顶间的交线和交点,并结合已得到的建筑边界、屋顶间交线和交点获取三维建筑模型.实验结果表明,该方法思路清晰、科学可行且适用性强,提取的建筑边界和模型垂直精度高.     

基于栅格填充的直角多边形建筑物轮廓规则化方法

智慧城市建设需要以城市全面感知与三维建模作为空间基础,而建筑物轮廓的自动提取和规则化则是建筑物三维建模的关键。现有的规则化方法由于只利用了轮廓的局部信息,存在拐点位置极难准确判断且容易漏判误判,规则化之后的轮廓并非规范正交等问题。提出一种基于栅格填充的规则化方法,利用轮廓的整体信息进行规则化,并利用图像处理中的腐蚀、膨胀算法进行优化,无需定位拐点且规则化结果规范正交。实验分析,该算法对噪声点的抑制力强,规则化结果与实际轮廓相似性高,尤其适用于从激光雷达扫描点云、倾斜摄影密集匹配点云中提取的直角多边形建筑物轮廓的规则化。     

基于Ly特征提取算子的高精度配准方法

本文探讨了一种基于Ly特征提取算子的遥感影像的高精度配准方法。该方法利用Ly算子快速提取特征点,同时采用抑制局部非最大的格网技术控制特征点均匀分布,在少量准确控制点的基础上,以相关系数测度及两个影像灰度值差的绝对值和测度相结合的多重判据为匹配准则,采用松弛法匹配策略整体匹配得到同名点。并通过实验证明了该方法特征提取速度快,基于Ly算子得到的配准点可靠,配准精度高。     

Derivation of a Tidal Inundation Model to Support Environmental Research in Roebuck Bay (Western Australia)

A method for deriving mean inundation times for a large expanse of intertidal mudflats is presented. Image processing techniques are used to extract waterlines from two Landsat Thematic Mapper scenes at approximately high and low tides, respectively. Along with data based on standard tidal predictions, the waterlines are utilised to derive an initial Digital Elevation Model (DEM) for the tidal flat within a GIS environment. As the model fails to fully cover the mean spring tidal range, a routine is developed to linearly extrapolate this surface to the elevation of the mean spring low tide. Tidal interpolation equations are then utilised to generate an inundation time in hours for each cell on this extrapolated DEM. Output from the generalised model is a series of maps showing isolines of inundation for both the mean spring and neap tidal cycles, as well as sets of point data that may be integrated with existing biological sample data. These data will be used to support an ongoing geologic and biologic investigation of the bay and its ecology.     

Recovery of elevation from estimated gradient fields constrained by digital elevation maps of lower lateral resolution

Depth measurement techniques like stereo analysis and laser range scanning often yield a lateral resolution below image resolution. In contrast, shading-based methods estimate the elevation model at image resolution. We present a computationally efficient approach to recover a surface of high vertical and high lateral resolution from a noisy gradient field and independently measured elevation data of lower lateral resolution, relying on a minimization of the mean squared difference between the low-pass component of the surface obtained based on shading information and that of the independently measured elevation data. The presented method is compared to a reference approach that minimizes a weighted sum comprised of the mean squared difference between the low-pass components of the estimated gradient field and the optimized model, respectively. The presented algorithm is applied using lunar orbital image data and stereo elevation data and is evaluated regarding orbital laser altimeter measurements of high vertical accuracy.     

A practical method for SRTM DEM correction over vegetated mountain areas

Digital elevation models (DEMs) are essential to various applications in topography, geomorphology, hydrology, and ecology. The Shuttle Radar Topographic Mission (SRTM) DEM data set is one of the most complete and most widely used DEM data sets; it provides accurate information on elevations over bare land areas. However, the accuracy of SRTM data over vegetated mountain areas is relatively low as a result of the high relief and the penetration limitation of the C-band used for obtaining global DEM products. The objective of this study is to assess the performance of SRTM DEMs and correct them over vegetated mountain areas with small-footprint airborne Light Detection and Ranging (Lidar) data, which can develop elevation products and vegetation products [e.g., vegetation height, Leaf Area Index (LAI)] of high accuracy. The assessing results show that SRTM elevations are systematically higher than those of the actual land surfaces over vegetated mountain areas. The mean difference between SRTM DEM and Lidar DEM increases with vegetation height, whereas the standard deviation of the difference increases with slope. To improve the accuracy of SRTM DEM over vegetated mountain areas, a regression model between the SRTM elevation bias and vegetation height, LAI, and slope was developed based on one control site. Without changing any coefficients, this model was proved to be applicable in all the nine study sites, which have various topography and vegetation conditions. The mean bias of the corrected SRTM DEM at the nine study sites using this model (absolute value) is 89% smaller than that of the original SRTM DEM, and the standard deviation of the corrected SRTM elevation bias is 11% smaller.     

基于重构等高线的DEM精度评估模型

利用重构等高线提出了一种评估数字高程模型精度的模型.基于该评估模型,数字高程模型的精度能利用提出的重构误差--等高线的偏移误差面积与原等高线的长度之比来评估.提出的重构误差具有惟一性,较之于常用的中误差具有更好的适宜性和可信度.     

全站扫描点云支持下的铁路隧道横断面快速提取及变形分析

使用全站扫描仪对采动区的铁路隧道进行扫描,在分析了测边后方交会设站扫描方法精度的基础上对拱形隧道的变形进行了研究,提出了通过拱形隧道的拱顶轴线完成横断面快速提取的方法。该方法基于高程差异获得隧道拱顶点集,而后采用随机采样一致算法对其处理并对处理后的点云拟合得到拱顶轴线,以该轴线方向作为横断面的法向完成横断面点集的快速提取,并基于所提取横断面点集获得横断面轮廓线并计算隧道变形。使用全站扫描仪的点云数据进行试验并进行变形分析。结果表明该方法能够完成对拱形铁路隧道横断面的快速提取,并且所提取的横断面能够反映隧道变形情况。     

ICESat-2/ATLAS全球高程控制点提取与分析

ICEsat-2/ATLAS是目前高程精度最高的星载激光数据,其数据覆盖全球,能够作为生产高精度全球地面参考高程的基础数据.基于ICESat-2/ATLAS全球激光数据产品ATL08,获取了全球ICESat-2陆地高程点,研究了基于参考高程数据和属性参数提取全球高程控制点的方法,并利用高精度参考高程数据验证了其精度.利...     

Multivariate Interpolation of Precipitation Using Regularized Spline with Tension

Regularized Spline with Tension (RST) is an accurate, flexible and efficient method for multivariate interpolation of scattered data. This study evaluates its capabilities to interpolate daily and annual mean precipitation in regions with complex terrain. Tension, smoothing and anisotropy parameters are optimized using the cross-validation technique. In addition, smoothing and rescaling of the third variable (elevation) is used to minimize the predictive error. The approach is applied to data sets from Switzerland and Slovakia and interpolation accuracy is compared to the results obtained by several other methods, expert-drawn maps and measured runoff. The results demonstrate that RST performs as well or better than the methods tested in the literature. The incorporation of terrain improves the spatial model of precipitation in terms of its predictive error, spatial pattern and water balance.     

球面小波模型在GPS高程拟合中的应用

本文利用某地共99个GPS水准点的大地高和正常高求取高程异常,使用DOG球面小波模型和多面函数,分别对高程异常进行拟合。拟合时剔除模型残差大于2倍中误差的点,并在剔除粗差后重新进行拟合。通过比较外部检核点的已知高程异常值和球面小波模型值、多面函数拟合值之间的均方差评价模型的精度。数据结果表明,以外部检核均方差最小为准则,球面小波模型拟合精度较优,其拟合精度为1.65 cm,多面函数拟合精度为2.35 cm。     

利用von Karman模型的Kriging插值用于GPS高程拟合

针对传统的GPS高程拟合方法要求学习样本分布要均匀的缺点,以平滑自协方差函数为基础,应用基于von Karman模型的Kriging方法用于GPS高程拟合,提出"曲线法"来计算von Karman模型中的v值。通过某测区GPS联测水准数据的计算,与多项式曲面拟合法及传统半变异函数Kriging插值等方法比较,表明:基于vonKarman模型的Kriging方法,对于学习样本数据分布不均的地区拟合精度更高,更适合地形条件恶劣地区GPS高程拟合。     

基于多方向剖面线Douglas-Pucker算法简化的DEM综合方法

文中提出一种基于多方向剖面线Douglas-Pucker(DP)算法简化的数字高程模型(DEM)综合方法,首先通过一定角度间隔旋转DEM,获得不同方向的剖面线;然后通过剖面线DP算法的简化获得特征;在此基础上,通过抑制局部极值,融合多方向的特征点,同时消除冗余;最后利用这些点构建三角网,再内插成规则DEM获得保持特征的DEM综合结果。利用两组2m格网的DEM数据进行实验,通过结果的定量和定性分析可以发现,文中方法比传统两个方向的简化方法可以获得更高精度的综合结果。