TUMTraf V2X Cooperative Perception Dataset

TUMTraf V2X Cooperative Perception Dataset是由慕尼黑工业大学创建的多模态数据集，专注于协同感知任务，包含2000个标记的点云和5000张图像，涵盖了从路边和车载传感器获取的数据。数据集包括30,000个带有跟踪ID的3D框，以及精确的GPS和IMU数据，适用于3D物体检测和跟踪任务。该数据集特别关注复杂的交通场景，如交通违规、近距离错过事件、超车和U型转弯，旨在通过协同感知技术提高自动驾驶车辆的安全性和可靠性。

The TUMTraf V2X Cooperative Perception Dataset is a multimodal dataset developed by the Technical University of Munich, dedicated to cooperative perception tasks. It encompasses 2000 annotated point clouds and 5000 images, with data collected from both roadside and on-board vehicle sensors. The dataset further includes 30,000 3D bounding boxes paired with tracking IDs, alongside precise GPS and IMU data, making it suitable for 3D object detection and tracking tasks. This dataset specifically focuses on complex traffic scenarios, such as traffic violations, near-miss incidents, overtaking, and U-turns, with the goal of enhancing the safety and reliability of autonomous vehicles via cooperative perception technologies.