一种从LiDAR点云中提取海岸线的新方法

针对利用等值线追踪法从LiDAR点云中提取的海岸线过于破碎、曲折和抖动大的问题, 提出了一种基于LiDAR点云数据栅格化的海岸线提取新方法。该方法通过剔除LiDAR点云中的粗差, 对LiDAR点云进行栅格化, 直接从栅格数据中提取出平均大潮高潮线。试验表明, 与等值线追踪法相比, 新方法简化了提取海岸线的过程, 提取的海岸线也更为平滑和合理。     

联合机载LiDAR数据和潮汐数据自动提取潮位线

首先,联合机载激光雷达(LiDAR)数据提取的海岸带数字表面模型(DSM)与验潮站数据提取的高、低潮面进行相交运算,生成"水陆二值图像",然后对其以提取边缘的方式提取高、低潮潮位线;针对LiDAR光束无法穿透水体导致低潮线附近DSM为无效值的缺陷,采取移动趋势面拟合法外推概略低潮线附近DSM,在此基础上重新提取更精确的低潮潮位线。实验表明,该方法能在较少人工干预的情况下有效提取高、低潮潮位线。     

利用 LiDAR 点云提取有地形约束的光滑海岸线

针对传统海岸线提取算法误差大、在地形突变区域形变大的缺点,提出了一种基于离散 LiDAR 点云提取有地形约束海岸线的方法.首先,通过离散点云构建约束三角网,减少内插规则格网的误差并且保证海岸线不穿越地形结构;其次,进行顾及地形结构的点云高程修正,在数据源上缓解海岸线的平滑问题;最后采取二次多项式法消除毛刺,利用动态阈值张力样条函数内插生成光滑海岸线.     

机载LiDAR点云数据的组合滤波算法研究

针对采用渐进式形态学滤波算法进行机载LiDAR点云滤波时存在的滤波效果不佳、地形特征保留不明显的问题,本文提出了一种改进不规则三角网的后处理滤波算法,构建组合式机载LiDAR点云滤波算法。该组合算法有效地结合了渐进式形态学滤波算法与改进TIN滤波算法的优势,首先采用渐进式形态学滤波算法对原始机载LiDAR点云数据进行处理,提取得到初始地面点;其次优化传统TIN滤波算法,以初始地面点及种子点构建TIN,通过连续迭代提取得到精细化地面点。为验证本文提出滤波算法的可靠性与优越性,选取宁波市某地2组机载LiDAR点云数据进行实验,结果表明,与较单一的渐进式形态学滤波算法、TIN滤波算法地面点提取结果相比较,本文改进滤波算法提取地面点的Ⅰ类误差、Ⅱ类误差及总误差均更低,且不受地形条件限制,具有较高的适应性,验证了本文提出改进滤波算法的可靠性与优越性。     

基于LiDAR的海岛岸线自动提取研究

为了探讨栅格化方法与图像分割法对海岛岸线提取的效果,本文主要研究了基于Li DAR数据利用这两种方法对某岛进行瞬时海岸线的提取。利用栅格化方法通过对Li DAR点云数据进行粗差剔除、滤波去噪、构建Terrain数据集、创建栅格表面、生成TIN模型及自动生成等高线,从而实现了瞬时海岸线的提取;利用图像分割法是通过对Li DAR点云数据进行粗差剔除、滤波处理、构建TIN模型、生成二值栅格图像、图像处理与图像边缘提取的过程实现瞬时海岸线的提取。对两种方法提取的海岸线进行叠加显示分析,试验结果表明:两种方法提取的海岸线形态结构基本吻合,海岸线提取效率较传统方法均有提高,但栅格化方法提取的海岸线比图像分割方法提取的海岸线更平滑、更细化,边缘信息较为丰富,与实际海岸线更贴切,效果更加理想。     

海岸线提取技术研究进展

探讨了海岸线定义不明确的问题。在总结实地测量痕迹线法和遥感影像提取法在测量海岸线的实现过程和关键技术的基础上,分析指出了这两种方法都存在效率低下、工作周期长、安全隐患大等缺点。介绍了国外利用航空LiDAR数据提取海岸线的主要方法及其实现过程,美国在垂直基准转换系统VDatum的支撑下,LiDAR技术提取海岸线法获得了较广泛的应用,而国内在技术方法研究与海域垂直基准建设方面则处于起步阶段。最后针对海岸线测量存在的问题提出了几点建议。     

海岸线提取技术研究进展

探讨了海岸线定义不明确的问题。在总结实地测量痕迹线法和遥感影像提取法在测量海岸线的实现过程和关键技术的基础上,分析指出了这两种方法都存在效率低下、工作周期长、安全隐患大等缺点。介绍了国外利用航空LiDAR数据提取海岸线的主要方法及其实现过程,美国在垂直基准转换系统VDatum的支撑下,LiDAR技术提取海岸线法获得了较广泛的应用,而国内在技术方法研究与海域垂直基准建设方面则处于起步阶段。最后针对海岸线测量存在的问题提出了几点建议。     

面向海洋应用的无人机遥感图像配准研究

在海洋应用中,大面积水体的同名点匹配相比陆地更加困难,制约了无人机遥感图像的配准精度和收敛速度。本文提出了一种改进算法适用于海洋无人机遥感应用,采用主成分分析（PCA）和水体阈值方法去除水体,获得图像中非水体区域的分块图像,然后利用仿射-尺度不变特征变换算法（ASIFT）进行图像的特征点提取和重叠图像非水体区域的同名点匹配。通过海岛、海岸线的无人机遥感试验结果表明,基于改进算法,在不增加时间开销的情况下,可以增加30%~50%的同名点数量,精度提高约5%~10%。文中方法适应用于海洋无人机遥感的序列图像配准,为海岛、海岸线的遥感监测提供了有效的技术支持。     

直线段检测算法在机载LiDAR建筑物轮廓线提取中的应用

针对机载LiDAR点云数据提取建筑物轮廓线耗时多且精度不高的问题,本文提出了一种基于直线段检测(LSD)的机载LiDAR建筑物轮廓线提取方法。该方法首先对已分类的建筑物点云进行栅格化得到二值图;然后对二值图进行膨胀、腐蚀操作,消除二值图中因栅格化产生的空洞;最后利用LSD算法进行直线检测获取规则的建筑物轮廓线。经过实测数据的验证,本文方法可以检测到亚像素级的建筑物轮廓线,与传统的Canny算法相比能够提高约50倍的效率。     

车载LiDAR点云中建筑物立面位置边界的自动提取

提出了一种基于点云特征图像和特征值分析的车载LiDAR点云建筑物立面位置边界的自动提取方法。首先利用车载LiDAR点云数据生成扫描区域的点云特征图像,并通过图像处理手段提取可能的建筑物目标点云;然后对提取的目标点云进行剖面分析和特征值分析,识别建筑物目标;最后对建筑物点云进行平面分割,提取建筑物立面,并对立面点云进行特征值分析,得到建筑物立面与地面交接的三维位置边界。实验结果表明,该方法能快速有效地提取车载LiDAR点云数据中的建筑物目标,同时提取的建筑物立面位置边界与原始点云能准确符合。     

支持向量机的建筑物激光脚点提取方法

本文提出了一种基于全色波段航空影像和激光雷达数据的建筑物检测方法.如何从激光点云数据中提取出建筑物激光脚点,是建筑物三维重建和轮廓提取的难点问题之一.植被密集区域以及与建筑物紧密相邻的树木的激光点很难与建筑物激光点区分开.本文利用支持向量机对单个激光点的特征进行两分类,特征向量包括激光点的高程、高程变化信息以及与激光点...     

利用LiDAR点云的真正射影像遮蔽检测

针对LiDAR系统真正射影像生成过程中由于高地物投影差产生的遮蔽问题,结合LiDAR点云数据提出了一种新的检测方法,将传统的基于格网的遮蔽区域检测转化为基于多尺度的TIN三角面片遮蔽检测。利用荆州市区获取的LiDAR数据对该方法与传统方法进行了比较,实验证明,该算法通过多重检测后忽略平坦地区,能够快速、准确地检测出遮挡区域。     

Detecting coastline change from satellite images based on beach slope estimation in a tidal flat

Beach heights and tidal variation have large impacts on the accuracy of estimates of coastline position and its historical changes of a wider and flatter beach based on remote sensing data. This study presents an approach to analysis of waterline movement based on the beach slope, estimated from two effective images with Landsat images data. Two images acquired at different stages of the tide were processed to delineate accurately the position of the waterline. Then waterlines were assigned heights using elevations predicted by a two-dimensional non-linear tidal assimilation model. Beach slope can be calculated piecewise using the heighted shorelines based on the equiangular triangle theory. The positions of the national tidal height datum coastline can be obtained by the beach slope calculation method to accurately monitor the changing of coastline. A change in the coastline of the southwest tidal flat of the Yellow River delta, from Tianshuigou to the Xiaoqing River mouth, was detected by combining field measurements of profiles and bathymetric data. The root mean squared error (RMSE) of the calculated slope of the intertidal zone was one order of magnitude less than the measured slope. The minimum error of self-consistency check is 0.2%. The RMSE between the coastlines estimated by the proposed method and those surveyed data varies from 53.98 m to 217.72 m. It is shown that this method is more suitable for the two years and over the time scales of shoreline change monitoring. To assess erosion/accretion patterns in the tidal flat, and the controlling factors, the volume of the beach was investigated as a possible indicator. The accepted coastline position and changes in the beach volume were used to monitor the changing pattern of accretion and erosion along the coast southwest of the recent Yellow River mouth.     

Automatic Registration Between Low-Altitude LiDAR Point Clouds and Aerial Images Using Road Features

Among the many means of acquiring surface information, low-altitude light detection and ranging (LiDAR) systems (e.g., unmanned aerial vehicle LiDAR, UAV-LiDAR) have become an important approach to accessing geospatial information. Considering the lower level of hardware technology in low-altitude LiDAR systems compared to that in airborne LiDAR, and the greater flexibility in-flight, registration procedures must be first performed to facilitate the fusion of laser point data and aerial images. The corner points and edges of buildings are frequently used for the automatic registration of aerial imagery with LiDAR data. Although aerial images and LiDAR data provide powerful support for building detection, adaptive edge detection for all types of building shapes is difficult. To deal with the weakness of building edge detection and reduce matching-related computation, the study presents a novel automatic registration method for aerial images, with LiDAR data, on the basis of main-road information in urban areas. Firstly, vector road centerlines are extracted from raw LiDAR data and then projected onto related aerial images with the use of coarse exterior orientation parameters (EOPs). Secondly, the corresponding image road features of each LiDAR vector road are determined using an improved total rectangle-matching approach. Finally, the endpoints of the conjugate road features obtained from the LiDAR data and aerial images are used as ground control points in space resection adjustment to refine the EOPs; an iterative strategy is used to obtain optimal matching results. Experimental results using road features verify the feasibility, robustness and accuracy of the proposed approach.     

Direct georeferencing of airborne LiDAR data in national coordinates

The topographic mapping products of airborne light detection and ranging (LiDAR) are usually required in the national coordinates (i.e., using the national datum and a conformal map projection). Since the spatial scale of the national datum is usually slightly different from the World Geodetic System 1984 (WGS 84) datum, and the map projection frame is not Cartesian, the georeferencing process in the national coordinates is inevitably affected by various geometric distortions. In this paper, all the major direct georeferencing distortion factors in the national coordinates, including one 3D scale distortion (the datum scale factor distortion), one height distortion (the earth curvature distortion), two length distortions (the horizontal-to-geodesic length distortion and the geodesic-to-projected length distortion), and three angle distortions (the skew-normal distortion, the normal-section-to-geodesic distortion, and the arc-to-chord distortion) are identified and demonstrated in detail; and high-precision map projection correction formulas are provided for the direct georeferencing of the airborne LiDAR data. Given the high computational complexity of the high-precision map projection correction approach, some more approximate correction formulas are also derived for the practical calculations. The simulated experiments show that the magnitude of the datum scale distortion can reach several centimeters to decimeters for the low (e.g., 500 m) and high (e.g., 8000 m) flying heights, and therefore it always needs to be corrected. Our proposed practical map projection correction approach has better accuracy than Legat’s approach,¹ but it needs 25% more computational cost. As the correction accuracy of Legat’s approach can meet the requirements of airborne LiDAR data with low and medium flight height (up to 3000 m above ground), our practical correction approach is more suitable to the high-altitude aerial imagery. The residuals of our proposed high-precision map projection correction approach are trivial even for the high flight height of 8000 m. It can be used for the theoretical applications such as the accurate evaluation of different GPS/INS attitude transformation methods to the national coordinates.     

基于LiDAR高度纹理和神经网络的地物分类

使用LiDAR单一数据进行点云分割工作时,基于斜率的严格分割LiDAR点云的方法不能很好的适应复杂地物 的分类工作。本文将LiDAR粗分割后的点云转换为高度图像和反射强度图像,并求取高度图像GLCM高度纹理。将4 种GLCM高度纹理、地面粗糙系数、平均高度和平均反射强度共7种纹理作为识别地面覆盖物的特征,并利用后向传播 神经网络(BP-ANN)方法对LiDAR数据进行地物识别。实验表明,这种方法能够从LiDAR独立数据源中有效的实现地 物分类,实验获得的精度大于90%。与传统的最大似然法进行对比,BP-ANN的分类精度高于最大似然法。当预设地 面类型能同时满足被光学影像和LiDAR数据识别的条件时,LiDAR高度纹理分类与光学影像分类结果的一致性达到 76.5%。     

分类线性回归的Landsat影像去云方法

首先,对参考影像和待去云影像上的云覆盖区域进行检测和掩膜;然后,对掩膜后的参考影像进行ISODATA聚类,并建立各个类别参考影像到待去云影像灰度值的线性回归方程;再对待去云影像上的云覆盖区域,依据参考影像上的灰度值进行最小距离方法分类,划分到聚类形成的各个类别之中;最后,依据各个类别回归方程进行灰度值预测。实验结果表明,所提方法能够进行云区的检测和去除,预测精度相比传统方法有较大提高。     

Extraction and motion estimation of vehicles in single-pass airborne LiDAR data towards urban traffic analysis

Airborne LiDAR data are characterized by involving not only rich spatial but also temporal information. It is possible to extract vehicles with motion artifacts from single-pass airborne LiDAR data, based on which the motion state and velocity of vehicles can be identified and derived. In this paper, a complete strategy for urban traffic analysis using airborne LiDAR data is presented. An adaptive 3D segmentation method is presented to facilitate the task of vehicle extraction. The method features an ability to detect local arbitrary modes at multi scales, thereby making it particularly appropriate for partitioning complex point cloud data. Vehicle objects are then extracted by a binary classification using object-based features. Furthermore, the motion analysis for extracted vehicles is performed to distinguish between moving and stationary ones. Finally, the velocity is estimated for moving vehicles. The applicability and efficiency of the presented strategy is demonstrated and evaluated on three ALS datasets acquired for the propose of city mapping, where up to 87% of vehicles have been extracted and up to 83% of moving traffic can be recovered together with reasonable velocity estimates. It can be concluded that airborne LiDAR data can provide value-added products for traffic monitoring applications, including vehicle counts, location and velocity, along with traditional products such as building models, DEMs and vegetation models.