Volumetric Semantically Consistent 3D Panoptic Mapping

We introduce an online 2D-to-3D semantic in- stance mapping algorithm aimed at generating comprehensive, accurate, and efficient semantic 3D maps suitable for au- tonomous agents in unstructured environments. The proposed approach is based on a Voxel-TSDF representation used in recent algorithms. It introduces novel ways of integrating semantic prediction confidence during mapping, producing semantic and instance-consistent 3D regions. Further improve- ments are achieved by graph optimization-based semantic labeling and instance refinement. The proposed method achieves accuracy superior to the state of the art on public large- scale datasets, improving on a number of widely used metrics. We also highlight a downfall in the evaluation of recent studies: using the ground truth trajectory as input instead of a SLAM-estimated one substantially affects the accuracy, creating a large gap between the reported results and the actual performance on real-world data. The code is available: https://github.com/y9miao/ConsistentPanopticSLAM.