From Fog to Failure: How Dehazing Can Harm Clear Image Object Detection

This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a multi-stage framework where a lightweight detector identifies regions of interest (RoIs), which are then enhanced via spatial attention-based dehazing before final detection by a heavier model. We analyze this phenomenon, investigate possible causes, and offer insights for designing hybrid pipelines that balance enhancement and detection. Our findings highlight the need for selective preprocessing and challenge assumptions about universal benefits from cascading transformations. Low-visibility conditions, such as rain, snow, fog, smoke, and haze, pose significant challenges for deep learning applications in autonomous vehicles, security and surveillance, maritime navigation, and agricultural robotics. Under these conditions, object detection models struggle due to reduced contrast and obscured features, leading to performance degradation. This study proposes a deep learning framework inspired by human visual perception to enhance object recognition in adverse visibility scenarios, particularly in foggy environments. A key motivation for this work comes from the impact of poor visibility on airport operations, where disruptions in taxiing and docking cause delays and increase reliance on ground support.