VeRi
收藏资源简介:
VeRi数据集是一个用于城市交通监控中车辆再识别的大型基准数据集。它包含超过50,000张776辆车的图像,这些图像由20个摄像头在24小时内覆盖1.0平方公里区域捕捉。数据集中的图像在真实世界不受约束的监控场景中捕捉,并标注了多种属性,如边界框、类型、颜色和品牌。每辆车被2至18个不同视角、光照、分辨率和遮挡情况的摄像头捕捉,提供了高重复率以适应实际监控环境中的车辆再识别。此外,数据集还标注了充足的牌照和时空信息,如牌照框、牌照字符串、车辆时间戳以及相邻摄像头之间的距离。
The VeRi dataset is a large-scale benchmark dataset designed for vehicle re-identification in urban traffic surveillance. It comprises over 50,000 images of 776 vehicles, captured by 20 cameras over a 24-hour period covering an area of 1.0 square kilometers. The images in the dataset were captured in real-world, unconstrained surveillance scenarios and are annotated with various attributes such as bounding boxes, type, color, and brand. Each vehicle is captured by 2 to 18 cameras with different perspectives, lighting conditions, resolutions, and occlusion levels, providing a high repetition rate to accommodate vehicle re-identification in actual surveillance environments. Additionally, the dataset is annotated with ample license plate and spatiotemporal information, including license plate bounding boxes, license plate strings, vehicle timestamps, and the distances between adjacent cameras.
VeRi数据集概述
数据集特性
- 规模与覆盖:包含超过50,000张图像,涉及776辆车辆,由20个摄像头在24小时内覆盖1.0平方公里区域。
- 场景与标注:图像在真实世界的非约束监控场景中捕获,并标注了多种属性,如边界框(BBoxes)、类型、颜色和品牌。
- 视角与条件:每辆车被2至18个不同视角、光照、分辨率和遮挡情况的摄像头捕捉。
- 额外信息:数据集还标注了丰富的车牌和时空信息,包括车牌边界框、车牌字符串、车辆时间戳及相邻摄像头间的距离。
数据集使用
- 获取方式:需通过电子邮件向联系人提供全名和单位信息,以确保数据集用于非商业目的。
引用信息
- 参考文献:使用该数据集时,应引用以下论文:
- Xinchen Liu, et al. "Large-scale vehicle re-identification in urban surveillance videos." ICME 2016.
- Xinchen Liu, et al. "A deep learning-based approach to progressive vehicle re-identification for urban surveillance." ECCV 2016.
- Xinchen Liu, et al. "PROVID: Progressive and multimodal vehicle reidentification for large-scale urban surveillance." IEEE Trans. Multimedia 20(3): 645-658 (2018).
性能评估
- 评估指标:数据集提供了多种评估指标,包括mAP、Rank-1、Rank-5等。
- 性能结果:展示了多个研究的年份和对应的性能结果,其中2019年的研究[22]在mAP上达到了最高值67.55%。




