Boxy

一个大型车辆检测数据集与几乎200万带注释的车辆训练和评估对象检测方法自动驾驶汽车在高速公路 ..
200,000图像1，990,000注释车辆500万像素分辨率阳光，雨，黄昏，夜间高速公路，交通繁忙，交通堵塞

This is a large-scale vehicle detection dataset designed for training and evaluating object detection algorithms for autonomous driving vehicles on highways, which contains nearly 2 million annotated vehicle samples. The dataset includes 200,000 images with a total of 1,990,000 annotated vehicle bounding boxes, all with a resolution of 5 megapixels. It covers various weather and time periods such as sunny, rainy, dusk and night, as well as multiple traffic conditions including highways, heavy traffic and traffic jams.

搜集汇总

背景与挑战

背景概述

Boxy是由Bosch Automated Driving于2019年发布的自动驾驶车辆检测数据集，包含20万张图像和199万车辆注释，覆盖多种天气和交通条件，图像分辨率为500万像素。