EvTurb: Event Camera Guided Turbulence Removal

2025-08-14

AI Summary

Atmospheric turbulence induces coupled blur and geometric distortion, severely degrading downstream vision tasks; this degradation is highly ill-posed. To address this, we propose the first event-camera-guided turbulence mitigation framework, leveraging the high temporal resolution of event streams to decouple and jointly correct both distortions: coarse deblurring is achieved via event integration, followed by variance-map-guided refinement of geometric correction. We introduce TurbEvent—the first real-scene, multi-scenario turbulent event dataset—and demonstrate state-of-the-art performance on both synthetic and real-world data, achieving superior restoration quality while maintaining real-time inference speed. Our core innovations include (i) an event-driven degradation decoupling model that explicitly separates blur and warping components, and (ii) a variance-map-guided fine-grained geometric correction mechanism that exploits spatiotemporal event statistics for robust alignment.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Time-Series/Data Streams

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Atmospheric turbulence degrades image quality by introducing blur and geometric tilt distortions, posing significant challenges to downstream computer vision tasks. Existing single-image and multi-frame methods struggle with the highly ill-posed nature of this problem due to the compositional complexity of turbulence-induced distortions. To address this, we propose EvTurb, an event guided turbulence removal framework that leverages high-speed event streams to decouple blur and tilt effects. EvTurb decouples blur and tilt effects by modeling event-based turbulence formation, specifically through a novel two-step event-guided network: event integrals are first employed to reduce blur in the coarse outputs. This is followed by employing a variance map, derived from raw event streams, to eliminate the tilt distortion for the refined outputs. Additionally, we present TurbEvent, the first real-captured dataset featuring diverse turbulence scenarios. Experimental results demonstrate that EvTurb surpasses state-of-the-art methods while maintaining computational efficiency.

Problem

• Removing blur and tilt distortions from turbulent images
• Decoupling turbulence effects using event camera data
• Creating a real-world turbulence dataset for evaluation

Innovation

• Event-guided turbulence removal framework
• Two-step event-guided network
• Real-captured TurbEvent dataset