Structure-Aware Radar-Camera Depth Estimation

TL;DR

This paper introduces a structure-aware radar-camera depth estimation framework that leverages RGB image priors and multi-scale guidance to produce dense, accurate depth maps, outperforming existing methods on the nuScenes dataset.

Contribution

It proposes a novel structure-aware strategy and a multi-scale network to enhance radar data processing for improved depth estimation.

Findings

Achieves state-of-the-art results on nuScenes dataset.

Effectively preserves detailed structures in depth maps.

Outperforms previous radar-camera depth estimation methods.

Abstract

Radar has gained much attention in autonomous driving due to its accessibility and robustness. However, its standalone application for depth perception is constrained by issues of sparsity and noise. Radar-camera depth estimation offers a more promising complementary solution. Despite significant progress, current approaches fail to produce satisfactory dense depth maps, due to the unsatisfactory processing of the sparse and noisy radar data. They constrain the regions of interest for radar points in rigid rectangular regions, which may introduce unexpected errors and confusions. To address these issues, we develop a structure-aware strategy for radar depth enhancement, which provides more targeted regions of interest by leveraging the structural priors of RGB images. Furthermore, we design a Multi-Scale Structure Guided Network to enhance radar features and preserve detailed structures, achieving accurate and structure-detailed dense metric depth estimation. Building on these, we propose a structure-aware radar-camera depth estimation framework, named SA-RCD. Extensive experiments demonstrate that our SA-RCD achieves state-of-the-art performance on the nuScenes dataset. Our code will be available at https://github.com/FreyZhangYeh/SA-RCD.