SynPick

SynPick是一个合成数据集，用于理解bin-pick场景中的动态场景。与现有数据集相比，该数据集位于现实世界的工业应用领域，并已获得ARC (Amazon Robotics Challenge) 的奖励，2017开发了具有逼真的拣选动作的动态场景。该数据集与流行的BOP数据集格式兼容。 数据集由21个合成视频组成，具有503,232个不同的照明，每个照明具有3个不同的视图。SynPick is a synthetic dataset dedicated to understanding dynamic scenes in bin-pick scenarios. Unlike existing datasets, it targets real-world industrial application scenarios and has been honored with awards from the ARC (Amazon Robotics Challenge). Developed in 2017, it features dynamic scenes with realistic picking motions. This dataset is compatible with the popular BOP dataset format. It consists of 21 synthetic videos, including 503,232 distinct lighting conditions, each paired with 3 different viewpoints.

搜集汇总

背景与挑战

背景概述

SynPick是一个用于理解动态bin-pick场景的合成数据集，专为工业应用设计，包含21个视频，涵盖超过50万种照明条件和多视图数据。该数据集由波恩大学于2021年发布，并与BOP数据集格式兼容。