查看文章

Recently, learning-based visual odometry (VO) has attained remarkable success in vision-based measurement, especially in indoor robotics. Unfortunately, existing methods usually underexplore geometric-semantic (G-S) information, thus resulting in inefficient perception in unseen dynamic environments. Meanwhile, they are usually time-consuming, since they typically rely on high-complexity semantic segmentation models, resulting in concurrence reduction. In this article, we develop a G-S information enhanced lightweight VO (GSL-VO) that can work particularly well in dynamic environments. Specifically, on the one hand, to improve the robustness of VO through G-S information, we first come up with a novel image enhancement module to tackle motion blur, thus enabling accurate geometric and semantic information extraction. Second, we design an adaptive G-S information processing module that combines …