DepthDark: Robust Monocular Depth Estimation for Low-Light Environments

2025-07-24

AI Summary

Monocular depth estimation foundation models suffer significant performance degradation under low-light conditions, primarily due to the absence of robust dark-light-specific architectures, large-scale high-quality paired depth datasets, and efficient fine-tuning paradigms. To address these challenges, we propose DepthDark: (1) We introduce the first high-fidelity nighttime paired depth dataset, uniquely integrating physics-driven glare and noise simulation modules to enhance realism and diversity. (2) We design a lighting-guided multi-scale feature fusion fine-tuning strategy, combined with parameter-efficient fine-tuning (PEFT) to improve generalization under illumination variation. Evaluated on nuScenes-Night and RobotCar-Night, DepthDark achieves state-of-the-art performance—demonstrating strong robustness with limited training data and computational resources. Our work establishes a novel paradigm for advancing foundation models in low-light monocular depth estimation.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In recent years, foundation models for monocular depth estimation have received increasing attention. Current methods mainly address typical daylight conditions, but their effectiveness notably decreases in low-light environments. There is a lack of robust foundational models for monocular depth estimation specifically designed for low-light scenarios. This largely stems from the absence of large-scale, high-quality paired depth datasets for low-light conditions and the effective parameter-efficient fine-tuning (PEFT) strategy. To address these challenges, we propose DepthDark, a robust foundation model for low-light monocular depth estimation. We first introduce a flare-simulation module and a noise-simulation module to accurately simulate the imaging process under nighttime conditions, producing high-quality paired depth datasets for low-light conditions. Additionally, we present an effective low-light PEFT strategy that utilizes illumination guidance and multiscale feature fusion to enhance the model's capability in low-light environments. Our method achieves state-of-the-art depth estimation performance on the challenging nuScenes-Night and RobotCar-Night datasets, validating its effectiveness using limited training data and computing resources.

Problem

• Lack robust monocular depth models for low-light conditions
• Missing high-quality paired depth datasets for nighttime scenarios
• Need effective parameter-efficient fine-tuning for low-light adaptation

Innovation

• Flare-simulation module for nighttime imaging
• Noise-simulation module for accurate datasets
• PEFT strategy with illumination guidance