【计算机科学】【2019.05】城市街道交叉口三维点云和照片模型的精度分析

云和照片模型的精度分析"/>

【计算机科学】【2019.05】城市街道交叉口三维点云和照片模型的精度分析

本文为美国南乔治亚大学（作者：Mariah D. Peart）的硕士论文，共156页。

南乔治亚大学建筑环境和建模实验室的这项研究比较了用最先进仪器和设备完成的点位置和距离测量。本研究使用了一台现代化的12秒激光扫描仪、一辆现代无人车和一台高精度的1秒机器人全站仪。后者作为基准工具。定量比较的主要目的是探索相对较大的点云模型的准确性和可用性，以作为重新设计/重建目的的虚拟测量工具。

该项目涉及生成两个繁忙、复杂的城市街道交叉口的大型三维点云模型。一个交叉口的面积约为300英尺×750英尺，包含五个汇合要素：三条街道和两条铁路。这是一个事故多发的地方，需要重新设计。第二个街道交叉口的面积约为1500英尺×2500英尺，包括两条以大约45度角相交的街道。由此产生的计算机模型已在乔治亚州东部州平面坐标系（SPCS）中进行地理参考，使用GPS（全球定位系统）通过快速、基于网络的实时运动学（RTK）方法建立坐标。这些城市街道交叉口位于乔治亚州斯泰茨伯罗市的Blue-Mile走廊内。Blue-Mile走廊与斯塔茨伯罗市的工程师们一道，计划加强和改善Blue-Mile走廊的交通流量，其中包括许多商业和餐馆。最终的点云模型将捐赠给城市工程师，以协助重新设计交叉口。文中对所提及的差异进行了全面分析，并提出了提高整体流量精度的建议。

From Georgia Southern University’s Built Environment and Modeling lab, this study compares point positions and distance measurements completed with state-of-the-art instruments and equipment. A modern, 12-second, laser scanner, a modern unmanned aerial vehicle and a highly accurate, 1-second robotic total station were employed for this study. The latter serving as the benchmark instrument. The main objective of this quantitative comparison is to explore the accuracy and usability of a relatively large point-cloud model, as a virtual surveying tool for redesign/reconstruction purposes. This project involves the generation of large, 3D, point-cloud models of two busy and complex city street intersections. One intersection encompasses an approximate area of 300 ft × 750 ft and contains five converging elements: three streets and two railroads. It is an accident-prone location requiring redesign. The second street intersection encompasses an approximate area of 1,500 ft × 2,500 ft, containing two streets intersecting at an approximate 45-degree angle. The resulting computer model has been geo-referenced in the Georgia East State Plane Coordinate System (SPCS) using control points with coordinates established by GPS (Global Positioning System) via a rapid, network-based, Real-Time Kinematic (RTK) approach. These city street intersections are within the Blue-Mile corridor in Statesboro, GA. Along with the Statesboro City Engineers, the Blue-Mile corridor has plans to enhance and improve the traffic flow of the Blue-Mile corridor, which contains many businesses and restaurants. The final point-cloud models are to be donated to the city engineers to assist in the redesign of the intersections. A full analysis of the referred discrepancies is presented and recommendations on improving the overall current accuracies are provided.

1. 引言
2. 技术背景
3. 采用的仪器与相关的性能
4. 案例研究1：地理参考、非地理参考和视觉对齐点云模型的精度
5. 案例研究2：点云模型与传统测量的精度
6. 案例研究3：点云和基于照片的模型与传统测量仪器的精度比较
7. 结果
8. 结论

更多精彩文章请关注公众号：