ConRebSeg
收藏资源简介:
ConRebSeg数据集由丹麦技术大学创建,专注于加固混凝土建筑的分割任务。该数据集包含14,805张RGB图像,涵盖了从自采集和YouTube视频中获取的多种喷射混凝土施工场景。数据集提供了四类物体的分割标签:暴露的钢筋、人员、汽车和卡车。创建过程中,数据集采用了序列化的方式,反映了机器人任务的时空特性。该数据集主要应用于建筑行业的自主机器人系统开发,旨在解决劳动力短缺和提高施工安全性的问题。
The ConRebSeg dataset, developed by the Technical University of Denmark, focuses on the segmentation task for reinforced concrete buildings. This dataset contains 14,805 RGB images, covering various shotcrete construction scenarios sourced from both self-collected footage and YouTube videos. It provides segmentation labels for four object categories: exposed rebar, personnel, cars, and trucks. During its creation, the dataset adopts a sequential format that reflects the spatiotemporal characteristics of robotic tasks. This dataset is primarily applied to the development of autonomous robotic systems in the construction industry, aiming to address labor shortages and enhance construction site safety.




