Waymo Open Dataset: Panoramic Video Panoptic Segmentation

Waymo Open Dataset: Panoramic Video Panoptic Segmentation是一个大规模数据集，旨在为自动驾驶提供高质量的全景分割标签。数据集由Waymo LLC创建，包含100,000张图像，分为2860个时间序列，每个序列由五个摄像头捕获，覆盖三个不同的地理位置。数据集的创建过程涉及从现有的Waymo Open Dataset中抽取图像，并为其添加时间上和跨摄像头一致的标签。该数据集的应用领域主要集中在自动驾驶技术，特别是在需要全面理解周围环境的场景中，如车辆导航和避障。

Waymo Open Dataset: Panoramic Video Panoptic Segmentation is a large-scale dataset designed to provide high-quality panoramic segmentation labels for autonomous driving. Developed by Waymo LLC, the dataset contains 100,000 images divided into 2,860 temporal sequences, each captured by five cameras and covering three distinct geographical locations. The dataset is constructed by extracting images from the existing Waymo Open Dataset and adding temporally and cross-camera consistent annotations to them. Its application fields mainly focus on autonomous driving technology, particularly in scenarios requiring comprehensive understanding of the surrounding environment, such as vehicle navigation and obstacle avoidance.