Motion blur compensated optical pose estimation for non-cooperative space targets using adaptive multi-scale ellipse...

Area of Science:

• Robotics and Autonomous Systems
• Computer Vision
• Aerospace Engineering

Background:

• Motion blur significantly degrades performance in optical navigation systems for non-cooperative spacecraft.
• Accurate relative pose estimation is critical for safe and successful spacecraft rendezvous, proximity operations, and docking.

Purpose of the Study:

• To develop an uncertainty-aware stereo imaging approach for robust motion blur restoration and relative pose recovery.
• To provide calibrated confidence in pose estimates for risk-aware spacecraft guidance and docking.

Main Methods:

• Coupling physics-guided deblurring with edge-preserving regularization for image restoration.
• Adaptive detection of elliptical features on manufactured surfaces to aid pose estimation.
• Stereo geometric refinement incorporating measurement uncertainty propagation.

Main Results:

• Demonstrated lower position and orientation errors compared to five baseline methods in laboratory and synthetic experiments.
• Achieved higher success rates, particularly under severe motion blur conditions.
• Maintained practical runtime for real-world applicability.

Conclusions:

• The proposed uncertainty-aware stereo imaging approach effectively mitigates motion blur in spacecraft navigation.
• Calibrated uncertainty estimates enable risk-aware decision-making for autonomous operations.
• The method offers significant improvements for challenging scenarios like proximity operations with non-cooperative targets.