Self-driving car dataset
收藏科学数据银行2025-08-25 更新2026-04-23 收录
下载链接：
https://www.scidb.cn/detail?dataSetId=f872f9c86f2f4a848f089e35f296e48d
下载链接
链接失效反馈官方服务：
资源简介：
为了使数据集尽可能多地包含复杂的道路场景，以满足算法对车辆遮挡行为和小目标检测的要求，解决现有数据集驾驶场景中检测目标类型单一、缺乏复杂场景的问题。通过现场采集，根据BDD100K和VOC2012数据集选取图像，收集并制作相应标签，得到10000张车视图图像的数据集，按照7：2：1的比例分为训练集、验证集和测试集。
To ensure the dataset covers as many complex road scenarios as possible, meet the algorithm's requirements for vehicle occlusion behavior detection and small object detection, and resolve the issues of limited detection target categories and lack of complex scenarios in current driving datasets. This dataset comprises 10,000 on-board view images, which were acquired via on-site collection. Images were selected based on the BDD100K and VOC2012 datasets, followed by the collection of valid image data and generation of corresponding labels. Finally, the dataset is split into training, validation and test sets at a ratio of 7:2:1.
提供机构：
高雨琦
创建时间：
2025-08-25
