top of page
Perception in Space
A Photorealistic Dataset and Vision-Based Algorithm for Anomaly Detection During Proximity Operations in Lunar Orbit
Selina Leveugle, Chang Won Lee, Svetlana Stolpner, Chris Langley, Paul Grouchy, Steven Waslander, Jonathan Kelly
IEEE Robotics and Automation Letters (RA-L), 2026.
​ALLO is a synthetic, photorealistic dataset designed to benchmark visual anomaly detection for spacecraft proximity operations in lunar orbit . Alongside the dataset, the paper introduces MRAD (Model Reference Anomaly Detection), a statistical algorithm that uses the known pose of Canadarm3 and a CAD model of the Gateway to generate reference images and flag deviations as anomalies.
FlowCLAS: Enhancing Normalizing Flow-Based Anomaly Segmentation Via Contrastive Learning
Chang Won Lee, Selina Leveugle, Paul Grouchy, Chris Langley, Svetlana Stolpner, Jonathan Kelly,
Steven L. Waslander
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2026.
​
Traditional normalizing flow-based anomaly segmentation methods underperform in complex, dynamic scenes. In FlowCLAS, we incorporate synthetic anomaly insertion as a data augmentation and add a contrastive loss to improve the discrimination between anomalies and inlier objects.
bottom of page