This is the official repository of the RadarOcc, a pioneering appraoch for 3D occupancy prediction based on 4D imaging radar. For technical details, please refer to our paper on NeurIPS 2024:
RadarOcc: Robust 3D Occupancy Prediction with 4D Imaging Radar
Fangqiang Ding1,*, Xiangyu Wen1,*, Yunzhou Zhu2, Yiming Li3, Chris Xiaoxuan Lu4,†
[arXiv] [page] [demo] [slide]
1University of Edinburgh, 2Georgia Institute of Technology, 3NYU, 4UCL
*Equal contribution, †Corresponding author
- [2024-05-22] Our preprint paper is available on 👉arXiv.
- [2024-09-26] Our paper is accepted by NeurIPS 2024 🎉.
- [2024-11-04] Our network and training code is uploaded. Stay tuned for update👀!
- [2024-11-11] Our demo video is available online. Watch it via 👉Youtube. GIFs are also provided 🔗below.
- [2024-11-12] Our paper 👉slides and recording have been uploaded to offical website.
- [2024-11-16] Our project page is published. Please have a look👉page.
- [TODO] Provide the illustration for dataset preparation and weight.
If you find our work helpful to your research, please consider citing:
@article{Ding_2024_NeurIPS,
title={Robust 3D Occupancy Prediction with 4D Imaging Radar},
author={Ding, Fangqiang and Wen, Xiangyu and Zhu, Yunzhou and and Li, Yiming and Lu, Chris Xiaoxuan},
journal={Advances in Neural Information Processing Systems (NeurIPS)},
year={2024}
}3D occupancy-based perception pipeline has significantly advanced autonomous driving by capturing detailed scene descriptions and demonstrating strong generalizability across various object categories and shapes. Current methods predominantly rely on LiDAR or camera inputs for 3D occupancy prediction. These methods are susceptible to adverse weather conditions, limiting the all-weather deployment of self-driving cars. To improve perception robustness, we leverage the recent advances in automotive radars and introduce a novel approach that utilizes 4D imaging radar sensors for 3D occupancy prediction. Our method, RadarOcc, circumvents the limitations of sparse radar point clouds by directly processing the 4D radar tensor, thus preserving essential scene details. RadarOcc innovatively addresses the challenges associated with the voluminous and noisy 4D radar data by employing Doppler bins descriptors, sidelobe-aware spatial sparsification, and range-wise self-attention mechanisms. To minimize the interpolation errors associated with direct coordinate transformations, we also devise a spherical-based feature encoding followed by spherical-to-Cartesian feature aggregation. We benchmark various baseline methods based on distinct modalities on the public K-Radar dataset. The results demonstrate RadarOcc’s state-of-the-art performance in radar-based 3D occupancy prediction and promising results even when compared with LiDARor camera-based methods. Additionally, we present qualitative evidence of the superior performance of 4D radar in adverse weather conditions and explore the impact of key pipeline components through ablation studies.
 |
|---|
Figure 1. Overall pipeline of RadarOcc. The data volume reduction pre-processes the 4DRT into a lightweight sparse RT via Doppler bins encoding and sidelobe-aware spatial sparifying. We apply spherical-based feature encoding on the sparse RT and aggregate the spherical features using Cartesian voxel queries. The 3D occupancy volume is finally output via 3D occupancy decoding. |
Here are some GIFs showing our qualitative results on 3D occupancy prediction. Foreground voxels are colored as red while background voxels are green. Some of these results can also be found in our supplementary demo video.
Please follow installation instructions from OpenOccupancy
Many thanks to these excellent projects: