SR-SLAM: Scene Reliability-Based RGB-D SLAM in Diverse Environments

Authors: Haolan Zhang, Chenghao Li, Thanh Nguyen Canh, Lijun Wang, and Nak Young Chong

Abstract: We propose SR-SLAM, a scene reliability-based RGB-D SLAM framework that achieves robust localization across diverse environments. The core innovation is a unified reliability assessment mechanism that integrates detection confidence, spatial distribution, feature quality, depth quality, and historical observations to comprehensively evaluate scene conditions. Based on this reliability assessment, the system adaptively guides decision-making throughout the entire SLAM pipeline, including dynamic region selection, feature removal strategy, pose refinement activation, keyframe selection criteria, and optimization weighting. Demonstrated superior robustness and accuracy improvement over state-of-the-art methods on TUM, BONN, and OpenLORIS-scene benchmarks, with validated real-world performance in various scenarios.

Recommended citation: H. Zhang, C. Li, T. N. Canh, L. Wang and N. Y. Chong, "SR-SLAM: Scene reliability-based RGB-D SLAM in diverse environments," Robotics and Autonomous Systems, vol. 197, 105306, 2026.
Download Paper