Dataset for Robust and Accurate Leading Vehicle Velocity Recognition

本数据集名为‘Dataset for Robust and Accurate Leading Vehicle Velocity Recognition’，由斯巴鲁公司与SIGNATE公司联合开发。数据集包含1000个场景，每个场景记录了100至200帧的图像数据，主要用于训练和评估领先车辆速度识别技术。数据集不仅涵盖正常驾驶环境，还包括夜间和雨天等对摄像头识别较为困难的环境。创建过程中，数据集通过商用现成系统（EyeSight）的摄像头采集图像，并确保数据量充足以支持机器学习。该数据集特别适用于开发车辆功能，如自动紧急制动和自适应巡航控制，旨在通过提供多样化的真实世界驾驶场景数据，推动自动驾驶技术的发展。

This dataset, named *Dataset for Robust and Accurate Leading Vehicle Velocity Recognition*, was co-developed by Subaru Corporation and SIGNATE. It includes 1000 driving scenarios, each containing 100 to 200 frames of image data, and is primarily used for training and evaluating leading vehicle velocity recognition technologies. The dataset covers not only standard driving environments but also challenging scenarios for camera-based recognition, such as nighttime and rainy conditions. During its development, images were captured using cameras from a commercial off-the-shelf system (EyeSight), with sufficient data volume ensured to support machine learning workflows. This dataset is particularly well-suited for developing automotive functions including automatic emergency braking and adaptive cruise control, and aims to advance autonomous driving technology by providing diverse real-world driving scenario data.