Vehicle-Rear
收藏资源简介:
Vehicle-Rear数据集是由巴西巴拉那联邦大学创建的,旨在通过非重叠摄像头进行车辆识别。该数据集包含超过三小时的高分辨率视频,详细记录了近3000辆车的品牌、型号、颜色和年份,以及车牌的位置和识别信息。数据集创建过程中,使用了低成本的500万像素CMOS图像传感器,同步时间,分辨率为1920x1080像素,每秒30.15帧。Vehicle-Rear数据集特别适用于探索车辆识别中的特征融合,尤其是在城市交通监控系统中,用于执法如速度限制执行、犯罪调查和交通流量分析。
The Vehicle-Rear dataset was developed by the Federal University of Paraná, Brazil, for vehicle recognition using non-overlapping cameras. This dataset contains over three hours of high-resolution videos, which comprehensively document the brand, model, color, production year of nearly 3000 vehicles, as well as the location and recognition information of their license plates. During the dataset creation process, low-cost 5-megapixel CMOS image sensors were employed, with synchronized timing, a resolution of 1920×1080 pixels, and a frame rate of 30.15 frames per second. The Vehicle-Rear dataset is particularly suitable for exploring feature fusion in vehicle recognition scenarios, especially in urban traffic monitoring systems for law enforcement applications such as speed limit enforcement, criminal investigations, and traffic flow analysis.




