PillarSegNet: Pillar-based Semantic Grid Map Estimation using Sparse LiDAR Data

TL;DR

PillarSegNet is a novel deep learning approach that converts sparse LiDAR point clouds into dense semantic grid maps using PointNet and 2D segmentation, improving urban environment understanding for autonomous vehicles.

Contribution

It introduces PillarSegNet, which learns features directly from 3D point clouds and produces dense semantic maps, outperforming previous grid map methods.

Findings

Achieves about 10% higher mIoU than state-of-the-art methods

Uses dense ground truth from multiple scans for training

Demonstrates effectiveness on the SemanticKITTI dataset

Abstract

Semantic understanding of the surrounding environment is essential for automated vehicles. The recent publication of the SemanticKITTI dataset stimulates the research on semantic segmentation of LiDAR point clouds in urban scenarios. While most existing approaches predict sparse pointwise semantic classes for the sparse input LiDAR scan, we propose PillarSegNet to be able to output a dense semantic grid map. In contrast to a previously proposed grid map method, PillarSegNet uses PointNet to learn features directly from the 3D point cloud and then conducts 2D semantic segmentation in the top view. To train and evaluate our approach, we use both sparse and dense ground truth, where the dense ground truth is obtained from multiple superimposed scans. Experimental results on the SemanticKITTI dataset show that PillarSegNet achieves a performance gain of about 10% mIoU over the state-of-the-art grid map method.