NYC-Indoor-VPR
收藏资源简介:
NYC-Indoor-VPR是由纽约大学创建的一个独特的室内视觉位置识别数据集,包含超过36,000张图像,这些图像来自纽约市13个不同的拥挤场景,拍摄条件多变,包括不同的光照和外观变化。数据集通过半自动标注方法建立地面实况,用于计算每张图像的位置信息。该数据集主要用于室内环境中的定位和导航研究,特别是在机器人和辅助导航系统中,旨在解决由于室内环境中的视觉重复和遮挡导致的定位难题。
NYC-Indoor-VPR is a unique indoor visual place recognition dataset created by New York University. It contains over 36,000 images captured from 13 distinct crowded indoor scenes across New York City, with variable capture conditions including different lighting and appearance changes. The dataset establishes ground truth using a semi-automatic annotation method to calculate the position information of each image. It is mainly used for research on localization and navigation in indoor environments, especially in robotics and assistive navigation systems, aiming to solve the localization challenges caused by visual repetitions and occlusions in indoor spaces.

- 1NYC-Indoor-VPR: A Long-Term Indoor Visual Place Recognition Dataset with Semi-Automatic Annotation纽约大学 · 2024年



