Gaussian Process Mapping of Uncertain Building Models With GMM as Prior

Mapping with uncertainty representation is required in many research domains, especially for localization. Although there are investigations regarding the of pose estimation an ego-robot map information, quality reference maps often neglected. To avoid potential problems caused by errors and a lack quantification, adequate measure required. In this paper, uncertain building models abstract surfaces using Gaussian Processes (GPs) proposed to describe probabilistic way. reduce redundant computation simple planar objects, extracted facets from Mixture Model (GMM) combined implicit GP map, also employing local GP-block techniques. The method evaluated on LiDAR point clouds city buildings collected mobile mapping system. Compared performance other methods such as Octomap, Occupancy Map (GPOM), Bayesian Generalized Kernel Inference (BGKOctomap), Local automatic relevance determination Hilbert (LARD-HM) Implicit Surface (GPIS), our achieves higher Precision-Recall AUC buildings.