IRIS-v2
收藏资源简介:
IRIS-v2是由巴黎萨克雷大学·波雷利中心与法国电力集团等机构联合创建的工业场景数据集,扩展自早期IRIS数据集。该数据集包含300张16384x8192分辨率的球形图像、530平方米的密集点云(150点/平方厘米)、约500个设备的功能示意图、6000个标注框和47000个分割掩码,数据通过LiDAR和Xphase Pro X2相机采集。数据集采用半自动方式构建CAD模型和3D管道路由信息,主要应用于数字孪生构建领域,旨在解决工业场景采集与功能示意图自动对齐的难题,为预测性维护和虚拟现实培训等应用提供支持。
IRIS-v2 is an industrial scene dataset jointly created by the Borelli Center of Paris-Saclay University, Électricité de France (EDF) and other institutions, which is expanded from the original IRIS dataset. This dataset contains 300 spherical images with a resolution of 16384×8192, dense point clouds covering 530 square meters (150 points per square centimeter), functional schematic diagrams of approximately 500 devices, 6000 annotation boxes, and 47000 segmentation masks. The data was collected using LiDAR and Xphase Pro X2 cameras. The dataset constructs CAD models and 3D pipeline routing information in a semi-automatic manner. It is mainly applied in the field of digital twin construction, aiming to solve the challenge of automatic alignment between industrial scene collected data and functional schematic diagrams, and provide support for applications such as predictive maintenance and virtual reality training.



