Globally-Stable and Robust Image-Based Visual Servoing for Positioning with Respect to a Cylinder
Alessandro Colotti, Francois Chaumette
AI summary
Problem
Traditional image-based visual servoing for cylindrical targets relies on estimating planar parameters from projected edges or requires visible cylinder ends, making it sensitive to noise, modeling errors, and impractical for long or unmarked pipelines. This limits reliable positioning for real-world applications like submarine inspection or aerial maneuvering.
Approach
The authors extract visual features directly from cylinder edge projections that map to the 3D pose, requiring only the cylinder radius. They design a controller using a generalized inverse to ensure global stability regardless of radius estimation accuracy, alongside an improved classical pseudo-inverse variant that simplifies interaction matrix computation.
Key results
- Proven global asymptotic stability independent of cylinder radius estimation accuracy
- Simplified interaction matrix estimation eliminating complex plane parameter dependencies
- Demonstrated improved convergence and robustness in simulations under severe modeling errors
- Validated real-world performance on a robotic arm for precise camera positioning
Why it matters
Enables reliable, calibration-robust visual servoing for cylindrical infrastructure and dynamic targets, expanding practical deployment in inspection and aerial robotics.
Abstract
This letter proposes a new image-based visual servo- ing controller for positioning a camera with respect to a cylin- drical object. Traditional image-based approaches often rely on estimating planar parameters from the cylinder’s projected edges, making them sensitive to noise and modeling errors. In this work, we introduce a novel controller that uses pure image features while directly tied to the cylinder’s 3D pose, which depends solely on the cylinder radius. Crucially, this controller offers formal global stability irrespective of the radius estimate. Simulations and real experiments with a robotic arm confirm the controller improved convergence and robustness under practical conditions.