ETHcavation
收藏资源简介:
ETHcavation数据集由苏黎世联邦理工学院的机器人系统实验室创建,专门用于动态施工环境中的全景场景理解和物体跟踪。该数据集包含502张手工标注的图像,涵盖35个语义类别,主要来源于施工现场。创建过程中,研究人员结合了2D全景分割和3D LiDAR数据,通过Kalman滤波进行动态物体检测和跟踪。该数据集的应用领域主要集中在自主导航和实时路径规划,旨在解决施工环境中动态、杂乱和非结构化环境的挑战。
The ETHcavation dataset was developed by the Robotics Systems Lab at ETH Zurich, and is specifically tailored for panoramic scene understanding and object tracking in dynamic construction environments. It comprises 502 manually annotated images spanning 35 semantic categories, which are primarily collected from construction sites. During its development, researchers integrated 2D panoramic segmentation and 3D LiDAR data, and employed Kalman filtering for dynamic object detection and tracking. The dataset is mainly applied in autonomous navigation and real-time path planning, aiming to address the challenges of dynamic, cluttered and unstructured construction environments.

- 1ETHcavation: A Dataset and Pipeline for Panoptic Scene Understanding and Object Tracking in Dynamic Construction Environments苏黎世联邦理工学院 · 2024年



