A Preprocessing and Postprocessing Voxel-based Method for LiDAR Semantic Segmentation Improvement in Long Distance

TL;DR

This paper introduces a preprocessing and postprocessing method for LiDAR semantic segmentation that enhances long-distance outdoor scene understanding by leveraging multi-scan data, significantly improving accuracy over existing single-scan approaches.

Contribution

The paper presents a novel multi-stage preprocessing and postprocessing approach that, combined with state-of-the-art models, improves long-distance LiDAR semantic segmentation performance.

Findings

Over 5% mIoU improvement in medium range

Over 10% mIoU improvement in far range

Effective in long-distance outdoor scenarios

Abstract

In recent years considerable research in LiDAR semantic segmentation was conducted, introducing several new state of the art models. However, most research focuses on single-scan point clouds, limiting performance especially in long distance outdoor scenarios, by omitting time-sequential information. Moreover, varying-density and occlusions constitute significant challenges in single-scan approaches. In this paper we propose a LiDAR point cloud preprocessing and postprocessing method. This multi-stage approach, in conjunction with state of the art models in a multi-scan setting, aims to solve those challenges. We demonstrate the benefits of our method through quantitative evaluation with the given models in single-scan settings. In particular, we achieve significant improvements in mIoU performance of over 5 percentage point in medium range and over 10 percentage point in far range. This is essential for 3D semantic scene understanding in long distance as well as for applications where offline processing is permissible.