Synthetic Dataset for Industrial Manipulation
收藏arXiv2024-03-31 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2404.00447v1
下载链接
链接失效反馈官方服务：
资源简介：
本研究创建的‘Synthetic Dataset for Industrial Manipulation’数据集，由卢森堡大学跨学科安全、可靠性和信任中心开发，专注于工业对象的自动化操作。数据集包含500张图像，每张图像均包含对象的姿态、场景图像及对象掩码，用于训练深度学习模型进行姿态估计。创建过程中，使用Blender软件基于CAD文件生成合成场景，并通过特殊着色技术模拟金属对象的真实外观。该数据集主要应用于工业机器人自动化操作，旨在通过精确的姿态估计提高机器人在非结构化环境中的灵活性和适应性。
The 'Synthetic Dataset for Industrial Manipulation' developed in this study was constructed by the Interdisciplinary Centre for Security, Reliability and Trust, University of Luxembourg, and focuses on automated manipulation of industrial objects. The dataset consists of 500 images, each containing the object's pose, scene image, and object mask, and is used to train deep learning models for pose estimation. During its development, Blender software was used to generate synthetic scenes based on CAD files, and special shading techniques were applied to simulate the realistic appearance of metallic objects. This dataset is primarily applied to industrial robotic automated manipulation, aiming to improve the flexibility and adaptability of robots in unstructured environments through accurate pose estimation.
提供机构：
卢森堡大学，跨学科安全、可靠性和信任中心（SnT）
创建时间：
2024-03-31
