D$^2$Turb: Depth-Aware Simulation and Decoupled Learning for Single-Frame Atmospheric Turbulence Mitigation

2026-05-25

AI Summary

Single-frame atmospheric turbulence restoration is highly ill-posed due to the coupling of spatially varying blur and non-rigid geometric distortions, making it challenging for existing methods to simultaneously recover texture details and correct geometry. This work proposes D²Turb, a unified framework that first introduces a depth-aware turbulence synthesis protocol to generate physically consistent degraded data and provide tilted supervision. It then employs a decoupled yet interactive two-stage restoration architecture, separating the task into texture deblurring and geometric correction, and integrates information across stages via an Adaptive Structural Prior Injection (ASPI) mechanism. By combining depth-aware phase-to-space modeling with dense optical flow-guided spatial unwarping, the method significantly outperforms current approaches on both synthetic and real-world data, achieving concurrent improvements in texture fidelity and geometric accuracy.

Abstract

Single-frame atmospheric turbulence mitigation is inherently ill-posed due to spatially varying blur coupled with non-rigid geometric distortion. Existing end-to-end approaches trained on flat-field simulations often struggle to balance texture recovery with geometric rectification. To overcome this limitation, we propose D$^2$Turb, a unified framework that bridges physics-grounded simulation with explicitly decoupled restoration. First, we introduce a Depth-Aware Turbulence Synthesis protocol that incorporates scene depth into the phase-to-space formulation. This generates physically consistent, depth-dependent degradations and provides a crucial intermediate tilt supervision signal for disentangled learning. Building upon this simulation engine, D$^2$Turb decomposes restoration into two interactive stages: texture deblurring and geometric rectification. The texture deblurring stage employs a deblurring backbone to recover fine-grained details while preserving geometric distortion for the subsequent rectification stage. To mitigate the information fragmentation commonly observed in cascaded designs, we further propose an Adaptive Structural Prior Injection (ASPI) mechanism that dynamically transfers deep structural representations from the deblurring module to guide dense flow prediction for spatial unwarping. Extensive experiments demonstrate that D$^2$Turb achieves state-of-the-art performance on both synthetic and real-world datasets, with consistent improvements in both texture recovery and geometric fidelity. Our code and pre-trained models are publicly available at https://github.com/HertzDot222/D2Turb.

Problem

• atmospheric turbulence mitigation
• single-frame restoration
• spatially varying blur
• geometric distortion
• ill-posed problem

Innovation

• Depth-Aware Simulation
• Decoupled Learning
• Atmospheric Turbulence Mitigation
• Adaptive Structural Prior Injection
• Single-Frame Restoration