Toronto-3D
收藏资源简介:
Toronto-3D是由加拿大滑铁卢大学地理与环境管理系创建的大型城市户外点云数据集,专门用于语义分割。该数据集通过移动激光扫描(MLS)系统在多伦多采集,覆盖约1公里道路,包含约7830万个点,分为8个标记对象类别。数据集的创建过程涉及使用32线激光雷达传感器、全景相机和全球导航卫星系统(GNSS)进行数据采集,并通过专业软件进行处理和标记。Toronto-3D数据集主要应用于自主驾驶和城市高清地图等领域,旨在通过提供高质量的标记数据,推动基于学习的深度学习模型的开发和测试。
Toronto-3D is a large-scale urban outdoor point cloud dataset developed by the Department of Geography and Environmental Management at the University of Waterloo, Canada, exclusively for semantic segmentation tasks. It was collected in Toronto via a mobile laser scanning (MLS) system, covering approximately 1 kilometer of road, containing around 78.3 million points, and categorized into 8 annotated object classes. The dataset creation process involved utilizing a 32-line LiDAR sensor, panoramic cameras, and a Global Navigation Satellite System (GNSS) for data acquisition, followed by data processing and annotation with professional software. The Toronto-3D dataset is mainly applied in domains such as autonomous driving and high-definition urban mapping, with the goal of promoting the development and testing of learning-based deep learning models by providing high-quality annotated data.

- 1Toronto-3D: A Large-scale Mobile LiDAR Dataset for Semantic Segmentation of Urban Roadways滑铁卢大学地理与环境管理系 · 2020年



