Data Generation for Learning to Grasp in a Bin-picking Scenario
This provides a valuable dataset for researchers in robotic grasping, though it is incremental as it builds on existing simulation and data generation methods.
The authors tackled the problem of generating large-scale grasping data for robotic bin-picking by creating a dataset of 100K samples in simulation using 77 objects from the YCB dataset, with varied environmental conditions like lighting and camera pose.
The rise of deep learning has greatly transformed the pipeline of robotic grasping from model-based approach to data-driven stream. Along this line, a large scale of grasping data either collected from simulation or from real world examples become extremely important. In this paper, we present our recent work on data generation in simulation for a bin-picking scene. 77 objects from the YCB object data sets are used to generate the dataset with PyBullet, where different environment conditions are taken into account including lighting, camera pose, sensor noise and so on. In all, 100K data samples are collected in terms of ground truth segmentation, RGB, 6D pose and point cloud. All the data examples including the source code are made available online.