Annotations for Pedestrian Environment (APE) dataset

APE数据集是由华盛顿大学西雅图分校创建的开放共享注释数据集，旨在学习城市行人路径网络。该数据集覆盖了北美和南美的6个城市，总面积约2700平方公里。数据集内容包括航空卫星图像、街道地图图像和栅格化注释，专注于行人路径环境，如人行道、过街通道和角落灯泡。创建过程涉及将GIS注释栅格化，以适应计算机视觉任务。APE数据集的应用领域包括自主驾驶、多模式导航、行程规划、移动模拟和货运管理，旨在解决城市规划系统和导航应用中行人路径信息缺失或不准确的问题。

The APE Dataset is an open-access annotated dataset created by the University of Washington, Seattle, aimed at learning urban pedestrian path networks. This dataset covers 6 cities across North and South America, with a total area of approximately 2700 square kilometers. Its contents include aerial satellite imagery, street map imagery, and rasterized annotations, focusing on pedestrian path environments such as sidewalks, crosswalks, and corner street lamps. The dataset creation process involves rasterizing GIS annotations to adapt to computer vision tasks. Application areas of the APE Dataset include autonomous driving, multi-modal navigation, trip planning, mobility simulation, and freight management, aiming to address the issue of missing or inaccurate pedestrian path information in urban planning systems and navigation applications.

搜集汇总

背景与挑战

背景概述

APE数据集是一个开放共享的行人环境注释数据集，由华盛顿大学西雅图分校创建，覆盖北美和南美6个城市，总面积约2700平方公里，提供航空卫星图像、街道地图图像和栅格化注释，专注于行人路径元素如人行道和过街通道。它通过栅格化GIS注释适应计算机视觉任务，应用于自主驾驶、多模式导航和城市规划等领域，旨在解决行人路径信息不准确的问题。