ABCDWaveNet: Advancing Robust Road Ponding Detection in Fog through Dynamic Frequency-Spatial Synergy

2025-04-07

AI Summary

To address the poor robustness of road puddle detection under foggy, low-visibility, and low-light conditions, this paper proposes a dynamic frequency-domain–spatial collaborative semantic segmentation framework. The method introduces adaptive dynamic convolution and a wavelet-enhanced module to enable joint frequency-spatial feature modeling, and incorporates an Adaptive Attention-Coupled Gating (AACG) mechanism to optimize multi-scale feature fusion. Furthermore, we construct Foggy Low-Light Puddle (FLLP), the first benchmark dataset specifically designed for puddle detection under foggy and weak-light conditions. Extensive experiments demonstrate state-of-the-art performance: the proposed method achieves IoU improvements of 3.51%, 1.75%, and 1.03% on Foggy-Puddle, Puddle-1000, and FLLP, respectively. On the Jetson AGX Orin platform, it attains a real-time inference speed of 25.48 FPS, satisfying the computational requirements of Advanced Driver Assistance Systems (ADAS).

Technology Category

• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Application Domains: Mobility, Driving & Flight

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Road ponding presents a significant threat to vehicle safety, particularly in adverse fog conditions, where reliable detection remains a persistent challenge for Advanced Driver Assistance Systems (ADAS). To address this, we propose ABCDWaveNet, a novel deep learning framework leveraging Dynamic Frequency-Spatial Synergy for robust ponding detection in fog. The core of ABCDWaveNet achieves this synergy by integrating dynamic convolution for adaptive feature extraction across varying visibilities with a wavelet-based module for synergistic frequency-spatial feature enhancement, significantly improving robustness against fog interference. Building on this foundation, ABCDWaveNet captures multi-scale structural and contextual information, subsequently employing an Adaptive Attention Coupling Gate (AACG) to adaptively fuse global and local features for enhanced accuracy. To facilitate realistic evaluations under combined adverse conditions, we introduce the Foggy Low-Light Puddle dataset. Extensive experiments demonstrate that ABCDWaveNet establishes new state-of-the-art performance, achieving significant Intersection over Union (IoU) gains of 3.51%, 1.75%, and 1.03% on the Foggy-Puddle, Puddle-1000, and our Foggy Low-Light Puddle datasets, respectively. Furthermore, its processing speed of 25.48 FPS on an NVIDIA Jetson AGX Orin confirms its suitability for ADAS deployment. These findings underscore the effectiveness of the proposed Dynamic Frequency-Spatial Synergy within ABCDWaveNet, offering valuable insights for developing proactive road safety solutions capable of operating reliably in challenging weather conditions.

Problem

• Detects road ponding robustly in foggy conditions
• Integrates dynamic frequency-spatial synergy for accuracy
• Improves ADAS performance in adverse weather scenarios

Innovation

• Dynamic convolution for adaptive feature extraction
• Wavelet-based frequency-spatial feature enhancement
• Adaptive Attention Coupling Gate for feature fusion