Enhanced GNSS/INS/Vision Navigation with FCN-Based Sky Segmentation for Robust Positioning in Urban Canyons

A novel tightly coupled GNSS/INS/Vision navigation system, named Sky-GVIO, is proposed to achieve continuous and accurate positioning in challenging urban canyon environments. The system leverages an adaptive sky-view image segmentation algorithm based on Fully Convolutional Networks (FCN) for robust GNSS non-line-of-sight (NLOS) detection and mitigation.