Baosheng Zhang

Papers

Total Citations

H-Index

About

Baosheng Zhang is a leading researcher in robust visual perception for autonomous mobile robots, with a primary focus on simultaneous localization and mapping (SLAM) and visual odometry (VO) in challenging, dynamic environments. His major contributions center on overcoming the limitations of traditional systems that assume static scenes. Zhang pioneered the development of feature-based stereo methods that remain reliable even when moving pedestrians and vehicles dominate the view. His most-cited work, "DynPL-SVO" (19 citations), introduces a robust stereo visual odometry that effectively filters dynamic points to maintain accurate motion estimation. He extended this concept in "DynPL-SLAM" (15 citations), a full SLAM system that leverages both points and line features for enhanced robustness in dynamic settings. Additionally, Zhang's "PC-IDN" (8 citations) presents a fast 3D loop closure detection method using projection context descriptors and incremental dynamic nodes, significantly improving long-term navigation accuracy. Through these innovations, Zhang has established himself as a key figure in advancing the real-world deployment of autonomous systems, directly addressing the critical challenge of operating safely among moving objects.