PLED-VINS: A Point-Line Event-Based Visual Inertial SLAM for Dynamic Environments

2026-07-08

AI Summary

This work addresses the degradation in visual SLAM accuracy caused by moving objects and rapid motion in dynamic environments by proposing a monocular event camera-based visual-inertial SLAM framework. The method introduces an entropy-proximity score map to assess the temporal reliability of point and line features, and integrates a unified robust point-line bundle adjustment with an adaptive weighting strategy to effectively suppress dynamic disturbances. Notably, it presents the first geometric reliability model for line features under motion conditions. Experimental results demonstrate that the proposed approach significantly improves state estimation accuracy on dynamic sequences containing moving objects, exhibiting superior robustness and effectiveness.

Abstract

Dynamic environments remain a fundamental challenge for visual SLAM, where unreliable observations from moving objects and rapid motion degrade state estimation accuracy. Although event cameras preserve fine-grained spatio-temporal information, most existing event-based SLAM frameworks still assume static scenes and lack approaches to estimate the reliability of features. To this end, we propose PLED-VINS, a monocular event camera-based visual-inertial SLAM framework that enables robust state estimation in dynamic environments. We propose an entropy-recency score map to characterize the temporal reliability of both point and line features based on event temporal statistics. Concurrently, geometric reliability is estimated via a unified point-line robust bundle adjustment. Building upon these, we design an adaptive weighting strategy that fuses temporal and geometric reliability, including motion-conditioned reliability modeling for line features, to suppress unreliable observations. Experimental results demonstrate that PLED-VINS improves state estimation on the evaluated dynamic sequences with moving objects.

Problem

• dynamic environments
• visual SLAM
• event camera
• state estimation
• unreliable observations

Innovation

• event-based SLAM
• dynamic environments
• point-line features
• temporal reliability
• robust bundle adjustment