Modeling evidential grids using semantic context information for dynamic scene perception

摘要: Abstract Uncertainty about urban environments stems not only from imprecise pose estimation and noisy information in images but also the lack of semantic information. This article presents an approach to improve perception capability intelligent vehicles complex environments. The new method uses meta-knowledge extracted context associated with depth model occupancy grids stereo vision. It evidential formalism Dempster–Shafer theory manage uncertainties involved grid discretization, partial observation environment dynamic elements present scene. Real experiments carried out a challenging using KITTI benchmark are reported, which meaningful evaluations compared standard done show that proposed is able better handle semantic, uncertainty aspects representation.