NYU-VPR
收藏资源简介:
NYU-VPR数据集是由纽约大学创建的一个大规模长期视觉位置识别基准数据集,包含超过200,000张图像,覆盖纽约大学校园附近2km×2km的区域,拍摄于2016年全年。该数据集特别关注图像视角方向和数据匿名化的影响,旨在解决自动驾驶车辆和视觉障碍人群辅助导航中的定位和地图构建问题。数据集通过车辆上的前视和侧视摄像头收集,具有GPS标签,并经过匿名化处理以保护隐私。该数据集的应用领域包括自动驾驶、辅助导航和城市环境中的长期位置识别研究。
The NYU-VPR Dataset is a large-scale long-term visual place recognition benchmark dataset developed by New York University. It contains over 200,000 images, covering a 2 km × 2 km area in the vicinity of the NYU campus, and was captured throughout the entire year of 2016. This dataset specifically focuses on the impacts of image viewing orientations and data anonymization, aiming to address localization and mapping problems in autonomous driving and assistive navigation for visually impaired people. Collected via front-facing and side-facing cameras mounted on a vehicle, the dataset is equipped with GPS labels and has undergone anonymization processing to protect user privacy. Its application fields include autonomous driving, assistive navigation, and long-term visual place recognition research in urban environments.

- 1NYU-VPR: Long-Term Visual Place Recognition Benchmark with View Direction and Data Anonymization Influences纽约大学 · 2022年



