CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Gaussian process mapping with uncertainty measures for the urban building models Authors:  Qianqian Zou - Leibniz University Hannover (Germany) [presenting]
Monika Sester - Leibniz University Hannover (Germany)
Abstract: Mapping with probabilistic uncertainty for urban scenes is required in many research domains, such as localization and sensor fusion. Although there are many uncertainty explorations in the pose estimation of an ego-robot with map information, the quality of the reference maps is often neglected. To avoid the potential problems caused by the errors of maps and a lack of uncertainty quantification, an adequate uncertainty measure for the maps is required. The modelling for urban buildings with the implicit surface using Gaussian Process (GP) is proposed to measure the mapping uncertainty in a probabilistic fashion. To reduce the redundant computation for simple planar objects, explicit facets from a Gaussian Mixture Model (GMM) are combined with the GP map while sparse GP techniques are used as well. The proposed method is evaluated on LiDAR point clouds of city buildings collected by a mobile mapping system. Compared to the performances of methods using standard GP, sparse GP and GMM, our method has shown lower RMSE and higher log-likelihood with less computational complexity.