Synthetic Dataset for Production Environment
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/DLR-RM/BlenderProc/
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是通过可扩展的流水线方法生成的综合数据集,旨在弥合生产环境中模拟与现实之间的差距,尤其适用于航空工业中的复杂装配。该数据集包含了在局部坐标框架中的网格和变换,其特性非常适合于物体检测。此外,该数据集还具备从CAD模型生成训练数据的可扩展流水线能力,适用于物体检测、分割以及姿态估计等任务。
This comprehensive dataset is generated through a scalable pipeline approach, designed to bridge the sim-to-real gap in production environments and is particularly suited for complex assembly tasks in the aerospace industry. It contains meshes and transforms defined within local coordinate frames, making it highly suitable for object detection applications. Furthermore, this dataset features a scalable pipeline for generating training data from CAD models, which is applicable to tasks including object detection, segmentation, and pose estimation.
提供机构:
Skotty research project



