Robust Monocular Depth Estimation under Challenging Conditions

TL;DR

This paper introduces md4all, a robust monocular depth estimation method that maintains high accuracy across various challenging conditions by leveraging complex sample generation and training strategies, outperforming prior approaches.

Contribution

The paper presents a novel training approach that enables monocular depth estimation models to perform reliably under adverse conditions without requiring modifications during inference.

Findings

Outperforms prior methods on nuScenes and Oxford RobotCar datasets.

Effective under both ideal and challenging illumination and weather conditions.

Achieves significant improvements in depth estimation accuracy.

Abstract

While state-of-the-art monocular depth estimation approaches achieve impressive results in ideal settings, they are highly unreliable under challenging illumination and weather conditions, such as at nighttime or in the presence of rain. In this paper, we uncover these safety-critical issues and tackle them with md4all: a simple and effective solution that works reliably under both adverse and ideal conditions, as well as for different types of learning supervision. We achieve this by exploiting the efficacy of existing methods under perfect settings. Therefore, we provide valid training signals independently of what is in the input. First, we generate a set of complex samples corresponding to the normal training ones. Then, we train the model by guiding its self- or full-supervision by feeding the generated samples and computing the standard losses on the corresponding original images. Doing so enables a single model to recover information across diverse conditions without modifications at inference time. Extensive experiments on two challenging public datasets, namely nuScenes and Oxford RobotCar, demonstrate the effectiveness of our techniques, outperforming prior works by a large margin in both standard and challenging conditions. Source code and data are available at: https://md4all.github.io.