NeRF-LiDAR: Generating Realistic LiDAR Point Clouds with Neural Radiance Fields

TL;DR

NeRF-LiDAR introduces a novel method for generating realistic LiDAR point clouds using neural radiance fields, leveraging real-world data to improve autonomous driving training and reduce labeling costs.

Contribution

This paper presents a new LiDAR simulation approach that uses real images and point clouds to learn scene representations, enhancing realism and training efficiency.

Findings

Trained segmentation models perform similarly on generated and real LiDAR data.

Generated data improves model accuracy through pre-training.

Method reduces dependence on expensive real labeled data.

Abstract

Labeling LiDAR point clouds for training autonomous driving is extremely expensive and difficult. LiDAR simulation aims at generating realistic LiDAR data with labels for training and verifying self-driving algorithms more efficiently. Recently, Neural Radiance Fields (NeRF) have been proposed for novel view synthesis using implicit reconstruction of 3D scenes. Inspired by this, we present NeRF-LIDAR, a novel LiDAR simulation method that leverages real-world information to generate realistic LIDAR point clouds. Different from existing LiDAR simulators, we use real images and point cloud data collected by self-driving cars to learn the 3D scene representation, point cloud generation and label rendering. We verify the effectiveness of our NeRF-LiDAR by training different 3D segmentation models on the generated LiDAR point clouds. It reveals that the trained models are able to achieve similar accuracy when compared with the same model trained on the real LiDAR data. Besides, the generated data is capable of boosting the accuracy through pre-training which helps reduce the requirements of the real labeled data.