SynPick: A Dataset for Dynamic Bin Picking Scene Understanding

TL;DR

SynPick is a synthetic dataset designed for dynamic scene understanding in bin-picking tasks, featuring realistic industrial scenarios, diverse actions, and pose tracking, to advance perception methods in robotics.

Contribution

The paper introduces SynPick, a novel synthetic dataset with dynamic scenes and realistic picking actions, tailored for industrial bin-picking perception research.

Findings

Pose tracking improves perception accuracy over single-shot estimation.

The dataset is compatible with BOP format and publicly available.

Baseline pose estimation results demonstrate the dataset's utility.

Abstract

We present SynPick, a synthetic dataset for dynamic scene understanding in bin-picking scenarios. In contrast to existing datasets, our dataset is both situated in a realistic industrial application domain -- inspired by the well-known Amazon Robotics Challenge (ARC) -- and features dynamic scenes with authentic picking actions as chosen by our picking heuristic developed for the ARC 2017. The dataset is compatible with the popular BOP dataset format. We describe the dataset generation process in detail, including object arrangement generation and manipulation simulation using the NVIDIA PhysX physics engine. To cover a large action space, we perform untargeted and targeted picking actions, as well as random moving actions. To establish a baseline for object perception, a state-of-the-art pose estimation approach is evaluated on the dataset. We demonstrate the usefulness of tracking poses during manipulation instead of single-shot estimation even with a naive filtering approach. The generator source code and dataset are publicly available.