TSDF++: A Multi-Object Formulation for Dynamic Object Tracking and Reconstruction

The ability to simultaneously track and reconstruct multiple objects moving in the scene is of the utmost importance for robotic tasks such as autonomous navigation and interaction. Virtually all of the previous attempts to map multiple dynamic objects have evolved to store individual objects in separate reconstruction volumes and track the relative pose between them. While simple and intuitive, such formulation does not scale well with respect to the number of objects in the scene and introduces the need for an explicit occlusion handling strategy. In contrast, we propose a map representation that allows maintaining a single volume for the entire scene and all the objects therein. To this end, we introduce a novel multi-object TSDF formulation that can encode multiple object surfaces at any given location in the map. In a multiple dynamic object tracking and reconstruction scenario, our representation allows maintaining accurate reconstruction of surfaces even while they become temporarily occluded by other objects moving in their proximity. We evaluate the proposed TSDF++ formulation on a public synthetic dataset and demonstrate its ability to preserve reconstructions of occluded surfaces when compared to the standard TSDF map representation. Code is available at https://github.com/ethz-asl/tsdf-plusplus.

Authors

• Margarita Grinvald
• Federico Tombari
• Roland Siegwart
• Juan I. Nieto 0001

Keywords

• Surface reconstruction
• Solid modeling
• Three-dimensional displays
• Shape
• Pipelines
• Cameras
• Robustness
• Object Tracking
• Dynamic Objects
• Dynamic Tracking
• Object Reconstruction
• Dynamic Reconstruction
• Truncated Signed Distance Function
• Dynamic Object Tracking
• Dynamic Object Reconstruction
• Multiple Objects
• Number Of Objects
• Local Map
• Reconstruction Accuracy
• Multiple Scenarios
• Individual Objects
• Object Surface
• Objects In The Scene
• Single Volume
• Map Representation
• Multiple Surface
• Object Pose
• Dense Reconstruction
• Object Instances
• Mask R-CNN
• Ray Casting
• Global Volume
• Implicit Representation
• Reconstruction Task
• Pose Tracking
• Object Volume