机载LiDAR树木检测在电力巡线中的应用研究

发布时间：2018-04-24 17:04

本文选题：LiDAR + 滤波分类　；参考：《兰州交通大学》2017年硕士论文

【摘要】：机载LiDAR测量技术是一种新兴的对地探测和定位模式,正广泛应用于铁路勘察设计、城市建模、森林资源管理、测绘等领域,尤其是数字化电力巡线中,引起了电网运营单位高度重视。数字化电网工程中,输电线路横跨区域较广,所处地形复杂,且导线长期暴露在野外,受到外界因素影响,会产生磨损、腐蚀等现象。因此,必须及时进行电力线路检测。传统巡线方法主要是人力,工作量大而且条件艰苦,尤其是山区和跨越江河的输电线路巡查,花费时间多、人力财力成本高、困难重重。而机载LiDAR技术以自身特点,很好地解决了这些难题,它通过获取电力线路走廊环境的三维可视化地理信息,可快速提取导线下方与地物点的距离,快速清查安全隐患点。在此背景下,本文基于机载LiDAR技术对梅州市平原县焦富甲乙线路进行树木隐患检测,结合激光脚点几何分布特征,以规则格网区域增长法为基础,应用冠层高度模型CHM与拉普拉斯算子研究树木提取,实现树障在架空输电线路中安全性分析,相关研究内容如下:1.通过阅读国内外近几年相关文献,深入研究机载LiDAR技术对地定位原理,深入掌握各传感器系统的实现机制,为集成系统提供了一定基础;结合实际经验,在惯性导航的数学基础和基本原理上推导出五种常用导航坐标系;总结出机载LiDAR数据采集预处理完整技术路线,为获取高精度点云数据质量做了准备。2.基于LiDAR点云数据无序不规则特性,实现点云数据高效组织管理。本文简要介绍了离散点K近邻查询层次包围盒与八叉树查询,采用格网分块K近邻查询对点云数据进行组织,极大的提高了数据处理效率,为后续点云滤波分类做了准备。3.基于点云空间分布模式对地物点的几何特征进行分析。树木和建筑物激光脚点在垂直方向和水平方向具有异同点,在DSM、DTM基础上利用区域连续性与非连续性或属性一致性,从提取的地物点中通过面积阈值大范围分离树木区和建筑物区,设置一定高程阈值,对树木区进行梯度分割求取拉普拉斯算子,再叠加规则化数字表面模型拉普拉斯算子得到树木,借助冠层高度模型,从而得到高等植被。4.基于提取的高等树木进行输电线路树障危险点预警研究分析。以架空输电线路运行规程提供的标准为参考依据,实现导线下方与地物点水平距离和垂直距离检测,判别危险点的实际位置信息。
[Abstract]:Airborne LiDAR survey technology is a new ground detection and positioning mode, which is widely used in railway survey and design, urban modeling, forest resource management, surveying and mapping, especially in the digital power line survey. Has caused the grid operation unit to attach great importance. In the digital power network project, the transmission lines span a wide area, the terrain is complex, and the conductors are exposed in the field for a long time, which will be affected by external factors, such as wear, corrosion and so on. Therefore, power line detection must be carried out in time. The traditional method of line patrol is mainly manpower, with heavy workload and difficult conditions. Especially, the inspection of transmission lines in mountainous areas and across rivers takes more time, high cost of human and financial resources, and many difficulties. The airborne LiDAR technology solves these problems well with its own characteristics. By obtaining 3D visual geographic information of the power line corridor environment, it can quickly extract the distance between the conductor and the ground point, and quickly check out the hidden safety points. Under this background, based on airborne LiDAR technology, the tree hidden danger detection of Jiaofujia and B lines in Pingyuan County of Meizhou City is carried out. Combined with the geometric distribution of laser foot points, the method of regular grid area growth is used as the basis. The canopy height model CHM and Laplace operator are used to study the tree extraction, and the safety analysis of tree barrier in overhead transmission line is realized. The related research contents are as follows: 1. Through reading the relevant literature at home and abroad in recent years, deeply studying the principle of airborne LiDAR technology to the ground location, deeply mastering the realization mechanism of each sensor system, providing a certain foundation for the integrated system, combining with the practical experience, Five commonly used navigation coordinate systems are derived from the mathematical basis and principle of inertial navigation, and the complete technical route of pre-processing of airborne LiDAR data acquisition is summarized, which is prepared to obtain the high accuracy point cloud data quality. 2. Based on the irregular character of LiDAR point cloud data, point cloud data organization and management are realized. In this paper, the hierarchical bounding box and octree query of discrete point K nearest neighbor query are briefly introduced, and the point cloud data is organized by using the grid block K nearest neighbor query, which greatly improves the data processing efficiency and prepares for the subsequent point cloud filtering classification. Based on the spatial distribution model of point cloud, the geometric characteristics of ground objects are analyzed. The laser foot-points of trees and buildings have similarities and differences in vertical and horizontal directions, using regional continuity and discontinuity or attribute consistency on the basis of DSM DTM. The tree area and the building area are separated from the extracted points by the area threshold, and a certain elevation threshold is set, and the Laplace operator is obtained by gradient segmentation of the tree area. Then the trees are obtained by superposing the regularized digital surface model Laplace operator and the canopy height model is used to obtain the higher vegetation. 4. Based on the extraction of higher trees, the early warning of obstacle point of transmission line tree is studied and analyzed. Based on the standard provided by overhead transmission line operation rules, the horizontal and vertical distance detection of the point below the conductor and the ground object is realized, and the actual position information of the dangerous point is distinguished.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM755

【参考文献】