Radar-Camera Pixel Depth Association for Depth Completion

TL;DR

This paper introduces a novel radar-to-pixel association method that improves depth completion by densifying radar returns and fusing radar with video data, outperforming previous methods on the nuScenes dataset.

Contribution

It proposes a new radar-to-pixel association stage that enhances depth completion by learning a mapping from radar returns to image pixels, addressing challenges of radar sparsity and wide beams.

Findings

Achieves superior depth completion performance on nuScenes dataset.

Effectively densifies radar returns for better fusion with video.

Outperforms camera and radar alone in depth estimation tasks.

Abstract

While radar and video data can be readily fused at the detection level, fusing them at the pixel level is potentially more beneficial. This is also more challenging in part due to the sparsity of radar, but also because automotive radar beams are much wider than a typical pixel combined with a large baseline between camera and radar, which results in poor association between radar pixels and color pixel. A consequence is that depth completion methods designed for LiDAR and video fare poorly for radar and video. Here we propose a radar-to-pixel association stage which learns a mapping from radar returns to pixels. This mapping also serves to densify radar returns. Using this as a first stage, followed by a more traditional depth completion method, we are able to achieve image-guided depth completion with radar and video. We demonstrate performance superior to camera and radar alone on the nuScenes dataset. Our source code is available at https://github.com/longyunf/rc-pda.