HazyDet: Open-source Benchmark for Drone-view Object Detection with Depth-cues in Hazy Scenes

2024-09-30

arXiv.org

AI Summary

To address the slow progress in object detection for foggy drone-captured imagery—largely due to the absence of a dedicated benchmark—this paper introduces HazyDet, the first open-source foggy drone detection benchmark comprising 383K real and synthetically fogged images. We further propose DeCoDet, a depth-aware detector designed specifically for this challenging domain. Its key contributions include: (1) a novel detection paradigm integrating depth priors; (2) a multi-scale depth-aware head coupled with a dynamic depth-conditioned convolution module; and (3) a scale-invariant reconstruction loss enabling pseudo-label–driven self-supervised learning of depth cues. Extensive experiments on HazyDet demonstrate that DeCoDet significantly outperforms state-of-the-art methods, especially for small and distant objects. All data, models, and code are publicly released to foster community advancement.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Hardware-aware ML
• Search and Optimization: Learning to Search

Application Category

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

Abstract

Drone-based object detection in adverse weather conditions is crucial for enhancing drones' environmental perception, yet it remains largely unexplored due to the lack of relevant benchmarks. To bridge this gap, we introduce HazyDet, a large-scale dataset tailored for drone-based object detection in hazy scenes. It encompasses 383,000 real-world instances, collected from both naturally hazy environments and normal scenes with synthetically imposed haze effects to simulate adverse weather conditions. By observing the significant variations in object scale and clarity under different depth and haze conditions, we designed a Depth Conditioned Detector (DeCoDet) to incorporate this prior knowledge. DeCoDet features a Multi-scale Depth-aware Detection Head that seamlessly integrates depth perception, with the resulting depth cues harnessed by a dynamic Depth Condition Kernel module. Furthermore, we propose a Scale Invariant Refurbishment Loss to facilitate the learning of robust depth cues from pseudo-labels. Extensive evaluations on the HazyDet dataset demonstrate the flexibility and effectiveness of our method, yielding significant performance improvements. Our dataset and toolkit are available at https://github.com/GrokCV/HazyDet.

Problem

• Object detection in hazy drone-view scenes lacks benchmarks
• Haze causes severe visual degradation in aerial images
• Existing methods struggle with synthetic-to-real domain shifts

Innovation

• Depth-Conditioned Kernel for dynamic feature modulation
• Progressive Domain Fine-Tuning for synthetic-to-real adaptation
• Scale-Invariant Refurbishment Loss for noisy depth learning