ROVR-Network/ROVR-Open-Dataset

ROVR开放数据集是一个旨在推动自动驾驶和机器人研究的数据集，它提供了丰富的真实世界数据，这些数据是从ADAS摄像头和LiDAR传感器中捕获的。该数据集覆盖了50多个国家，超过200万公里的驾驶数据，适合用于训练和开发用于深度估计、目标检测和语义分割的高级AI算法。数据集包括LiDAR点云、单目深度估计、目标检测和语义分割等多种数据类型。第一批ROVR开放数据集包括1363个视频片段，每个片段包含30秒数据，来自全球不同的环境，以支持自动驾驶研究。数据集分为高速公路、城市和乡村三种场景类型，并包括白天、夜晚和雨天等不同条件。数据集划分为训练集和测试集，当前提供单目深度估计的地面真实注释，计划在未来的版本中包括目标检测和语义分割的注释。

The ROVR Open Dataset is designed to empower autonomous driving and robotics research by providing rich, real-world data captured from ADAS cameras and LiDAR sensors. The dataset spans over 50 countries with more than 20 million kilometers of driving data, making it ideal for training and developing advanced AI algorithms for depth estimation, object detection, and semantic segmentation. The dataset includes various data types such as LiDAR point clouds, monocular depth estimation, object detection, and semantic segmentation. The first batch of the ROVR Open Dataset includes 1,363 clips, each containing 30 seconds of data, collected from diverse global environments to support autonomous driving research. The dataset covers three scene types - Highway, Urban, and Rural, with different conditions such as Day, Night, and Rainy. The dataset is split into a Training Set and a Test Set, with current ground truth annotations for monocular depth estimation and planned annotations for object detection and semantic segmentation in future releases.