Semantic Segmentation Based Rain and Fog Filtering Only by LiDAR Point Clouds

Zhen Luo, Junyi Ma, Guangming Xiong*, Xiuzhong Hu, Zijie Zhou, Jiahui Xu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

The basis of the autonomy of an intelligent vehicle is that hardware can provide reliable perceptual information. To apply the intelligent vehicles to the field of transportation, a problem that has to be solved is the autonomous driving in adverse weather scenes. Single categories of sensors, such as LiDAR, are often affected by adverse weather, which has led to the development of multi-sensor fusion technology, but has also resulted in increased costs. In this paper, we propose a point clouds denoising method based on semantic segmentation, and advance a post-processing method to improve the performance of the network. We implement a set of software packages under a ROS framework that only needs LiDAR to denoise in adverse weather. The experimental results show that our proposed method outperforms the existing mainstream methods in terms of filtering out rain and fog point clouds and the performance has been improved by 4.1% on MIoU.

Original languageEnglish
Title of host publicationProceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022
EditorsRong Song
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages90-95
Number of pages6
ISBN (Electronic)9781665484565
DOIs
Publication statusPublished - 2022
Event2022 IEEE International Conference on Unmanned Systems, ICUS 2022 - Guangzhou, China
Duration: 28 Oct 202230 Oct 2022

Publication series

NameProceedings of 2022 IEEE International Conference on Unmanned Systems, ICUS 2022

Conference

Conference2022 IEEE International Conference on Unmanned Systems, ICUS 2022
Country/TerritoryChina
CityGuangzhou
Period28/10/2230/10/22

Keywords

  • adverse weather
  • point clouds
  • rain and fog denoising
  • semantic segmentation

Fingerprint

Dive into the research topics of 'Semantic Segmentation Based Rain and Fog Filtering Only by LiDAR Point Clouds'. Together they form a unique fingerprint.

Cite this