POCD Pipeline

Our object-aware map update framework for semi-static environments. The system takes in RGB-D frames, each one semantically annotated and converted to a semantic point cloud. The point cloud is clustered into observations, and associated to mapped objects. A TSDF-based object-level change estimation is performed between the associated observation-object pairs, followed by a joint probabilistic update of the geometric change and stationarity score. Objects with a high stationarity score are used to generate the new map, and the low-score objects are discarded.

Abstract

Maintaining an up-to-date map to reflect recent changes in the scene is very important, particularly in situations involving repeated traversals by a robot operating in an environment over an extended period. Undetected changes may cause a deterioration in map quality, leading to poor localization, inefficient operations, and lost robots. Volumetric methods, such as truncated signed distance functions (TSDFs), have quickly gained traction due to their real-time production of a dense and detailed map, though map updating in scenes that change over time remains a challenge. We propose a framework that introduces a novel probabilistic object state representation to track object pose changes in semi-static scenes. The representation jointly models a stationarity score and a TSDF change measure for each object. A Bayesian update rule that incorporates both geometric and semantic information is derived to achieve consistent online map maintenance. To extensively evaluate our approach alongside the state-of-the-art, we release a novel real-world dataset in a warehouse environment. We also evaluate on the public ToyCar dataset. Our method outperforms state-of-the-art methods on the reconstruction quality of semi-static environments. The dataset is available here.

Paper

POCD: Probabilistic Object-Level Change Detection and Volumetric Mapping in Semi-Static Scenes
Jingxing Qian, Veronica Chatrath, Jun Yang, James Servos, Angela P. Schoellig, and Steven L. Waslander
Robotics: Science and Systems (RSS) 2022

@INPROCEEDINGS{QianChatrathPOCD,
  author={Qian, Jingxing and Chatrath, Veronica and Yang, Jun and Servos, James and Schoellig, Angela and Waslander, Steven L.},
  booktitle={2022 Robotics: Science and Systems (RSS)}, 
  title=, 
  year={2022},
  volume={},
  number={},
  pages={},
  doi={}}

Affiliations

Snow
Forest