GPSlam: Marrying Sparse Geometric and Dense Probabilistic Visual Mapping

We propose a novel, hybrid SLAM system to construct a dense occupancy grid map based on sparse visual features and dense depth information. While previous approaches deemed the occupancy grid usable only in 2D mapping, and in combination with a probabilistic approach, we show that geometric SLAM can produce consistent, robust and dense occupancy information, and maintain it even during erroneous exploration and loop closure. We require only a single hypothesis of the occupancy
 map and employ a weighted inverse mapping scheme to align it to sparse geometric information. We propose a novel map-update criterion to prevent inconsistencies, and a robust measure to discriminate exploration from localization.