Always Clear Depth: Robust Monocular Depth Estimation under Adverse Weather

TL;DR

This paper introduces ACDepth, a robust monocular depth estimation method that uses diffusion models, domain adaptation, and knowledge distillation to improve performance under adverse weather conditions.

Contribution

The paper proposes a novel approach combining diffusion-based data augmentation, fine-tuning with LoRA, and multi-granularity knowledge distillation for robust depth estimation in adverse weather.

Findings

ACDepth outperforms previous methods on nuScenes in night and rainy scenes.

The diffusion model effectively generates realistic adverse weather samples.

Knowledge distillation improves the model's degradation-agnostic scene understanding.

Abstract

Monocular depth estimation is critical for applications such as autonomous driving and scene reconstruction. While existing methods perform well under normal scenarios, their performance declines in adverse weather, due to challenging domain shifts and difficulties in extracting scene information. To address this issue, we present a robust monocular depth estimation method called \textbf{ACDepth} from the perspective of high-quality training data generation and domain adaptation. Specifically, we introduce a one-step diffusion model for generating samples that simulate adverse weather conditions, constructing a multi-tuple degradation dataset during training. To ensure the quality of the generated degradation samples, we employ LoRA adapters to fine-tune the generation weights of diffusion model. Additionally, we integrate circular consistency loss and adversarial training to guarantee the fidelity and naturalness of the scene contents. Furthermore, we elaborate on a multi-granularity knowledge distillation strategy (MKD) that encourages the student network to absorb knowledge from both the teacher model and pretrained Depth Anything V2. This strategy guides the student model in learning degradation-agnostic scene information from various degradation inputs. In particular, we introduce an ordinal guidance distillation mechanism (OGD) that encourages the network to focus on uncertain regions through differential ranking, leading to a more precise depth estimation. Experimental results demonstrate that our ACDepth surpasses md4all-DD by 2.50\% for night scene and 2.61\% for rainy scene on the nuScenes dataset in terms of the absRel metric.