IDD-3D

IDD-3D数据集是由海得拉巴国际信息技术研究所创建，专注于印度复杂非结构化道路场景的自动驾驶研究。该数据集包含来自多个摄像头和激光雷达传感器的多模态数据，总计12,000个标注的驾驶激光雷达帧，覆盖了多种交通场景。数据集的创建旨在解决现有数据集在地理多样性上的不足，特别是在发展中国家如印度的复杂交通布局。IDD-3D数据集通过提供高复杂度的轨迹和详细的3D对象边界框标注，支持了3D对象检测和跟踪任务的研究，适用于自动驾驶系统在复杂环境中的应用。

The IDD-3D dataset was created by the International Institute of Information Technology, Hyderabad, focusing on autonomous driving research in complex unstructured road scenarios across India. This dataset collects multimodal data from multiple cameras and LiDAR sensors, comprising a total of 12,000 annotated driving LiDAR frames that cover a wide range of traffic scenarios. The development of IDD-3D aims to mitigate the limitations of existing datasets in terms of geographic diversity, especially concerning the complex traffic configurations in developing nations such as India. By offering highly complex driving trajectories and precise 3D object bounding box annotations, the IDD-3D dataset supports research on 3D object detection and tracking tasks, and is well-suited for advancing autonomous driving systems in complex real-world environments.