Author(s)

Haifei Peng

Corresponding Author

Haifei Peng

Abstract

Accurate mapping of complex terrain is very important in the fields of geological disaster monitoring and smart city construction. However, traditional surveying and mapping methods are inefficient, costly and have incomplete data coverage in complex scenes. In this paper, a high-precision mapping method of complex terrain is proposed, which combines Unmanned Aerial Vehicle (UAV) remote sensing and 3D laser scanning. Through the collaborative observation scheme of "wide area coverage in the air+ground detail enhancement", combined with deep learning assisted point cloud registration and noise filtering technology, the measurement of plane accuracy ± 5cm and elevation accuracy ± 3cm has been achieved. In addition, the terrain adaptive filtering algorithm developed by the research has improved the completeness rate of extracting surface information from vegetation covered areas to over 90%. The effectiveness and universality of the method have been verified in typical scenarios, providing a new technical solution for high-precision mapping of complex terrains.

Keywords

3D laser scanning, high-precision mapping, complex terrain, UAV remote sensing