期刊
JOURNAL OF FIELD ROBOTICS
卷 33, 期 1, 页码 103-132出版社
WILEY
DOI: 10.1002/rob.21614
关键词
-
类别
资金
- Dr.-Ing.Willy-Hofler-Foundation scholarship grant
The ability to generate accurate and detailed three-dimensional (3D) maps of a scene from a mobile platform is an essential technology for a wide variety of applications from robotic navigation to geological surveying. In many instances, the best vantage point is from above, and as a result, there is a growing demand for low-altitude mapping solutions from micro aerial vehicles such as small quadcopters. Existing lidar-based 3D airborne mapping solutions rely on GPS/INS solutions for positioning, or focus on producing relatively low-fidelity or locally focused maps for the purposes of autonomous navigation. We have developed a general-purpose airborne 3D mapping system capable of continuously scanning the environment during flight to produce accurate and dense point clouds without the need for a separate positioning system. A key feature of the system is a novel passively driven mechanism to rotate a lightweight 2D laser scanner using the rotor downdraft from a quadcopter. The data generated from the spinning laser is input into a continuous-time simultaneous localization and mapping (SLAM) solution to produce an accurate 6 degree-of-freedom trajectory estimate and a 3D point cloud map. Extensive results are presented illustrating the versatility of the platform in a variety of environments including forests, caves, mines, heritage sites, and industrial facilities. Comparison with conventional surveying methods and equipment demonstrates the high accuracy and precision of the proposed solution.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据