Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Ronghui Zhang, Dakang Lyu, Tengfei Li, Yunfan Wu, Ujjal Manandhar, Benfei Wang, Junzhou Chen, Bolin Gao, Danwei Wang, Yiqiu Tan

TL;DR

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.

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.

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

TL;DR

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.

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.

Paper Structure

This paper contains 44 sections, 34 equations, 9 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Sensor-based methods
  4. Machine learning-based Methods
  5. Classical Machine Learning Methods
  6. Deep Learning-based Methods
  7. Foggy Conditions Detection
  8. Proposed Method
  9. Overall Architecture
  10. Dynamic Feature Extraction
  11. Dynamic Convolution
  12. Dual Dynamic Convolution Layer
  13. Wavelet-Based Synergy of Frequency and Spatial Information
  14. Frequency-Aware Information Processing Pathway
  15. Spatial-Enhanced Feature Learning Pathway
  16. ...and 29 more sections

Figures (9)

  • Figure 1: Road Ponding Detection in an Advanced Driver Assistance System (ADAS) zhaonan.
  • Figure 2: The Architecture of AGSENet(a): A U-Shaped Network Featuring Five Encoders, Four Decoders, Multi-Scale Information Aggregation (MIA), and Adaptive Attention Coupling Gate (AACG) Modules (d). The Encoder-Decoder Structure is Built by Stacking Dual Dynamic Convolution-Bidomain Information Synergy (DDC-BIS) Modules (b), with Each DDC Layer Comprising Two Consecutive Dynamic Convolution Layers (c).
  • Figure 3: Wavelet Sub-band Analysis of Puddle-1000 Images and Synthesized Foggy-Puddle Images.
  • Figure 4: (a) The architecture of Feature Mixing Block (FMBlock). (b) The architecture of Feature Mixing Block Conv (FMBconv). (c) The architecture of Contrast- Aware Channel Attention (CCA).
  • Figure 5: The Architecture of Multi-Scale Information Aggregation module. (a) illustrates the oversall computational process of MIA module. (b) details the computational specifics of the Adaptive Scale Selection (ASS). (c) shows the Progressive Separable Refinement module.
  • ...and 4 more figures