Wavelength Selection Method of Unmanned Vehicle-Mounted Multi-Wavelength Lidar in Off-Road Environment

Xia Yu-hao,Xu Shi-long,Hu Yi-hua
DOI: https://doi.org/10.1117/12.2604872
2021-01-01
Abstract:At present, the technical research of lidar used in unmanned vehicle driving mainly focuses on continuously improving the density of lidar point cloud under the working mode of lidar with single wavelength, but the detection of echo is limited to single echo, missing a lot of details. Although the increase of laser point cloud density can improve the object recognition ability based on the geometric features of the point cloud, it also has a decreasing effect and many additional system requirements, which cannot fundamentally solve the problem of the lack of physical property detection ability caused by the single wavelength of lidar. To promote cross-country environment physical properties of the laser radar detection ability and help the laser radar's ability to obtain the information such as target state, in this paper, based on the calculation results of typical target spectral characteristics and lidar echo characteristics, a wavelength selection method for unmanned multi-wavelength lidar in off-road environment is proposed, which uses principal component analysis of typical target spectral features to determine the characteristic wavelengths that can distinguish the target by spectral features. Besides, the degree of waveform splitting is discussed through the simulation calculation of laser echo waveform, which helps finding the spectral wavelengths to distinguish targets in the same distance.
What problem does this paper attempt to address?