PLT-D3

PLT-D3数据集由Pilotier多伦多创建，专注于提升自动驾驶系统在多样天气条件下的性能。该数据集包含高保真立体深度和场景流真实数据，使用Unreal Engine 5生成，涵盖雨、雪、雾及多种光照条件。数据集内容包括同步的高分辨率立体图像序列，旨在解决现有数据集在复杂天气模拟方面的不足。创建过程中，利用UE5的强大功能，实现了高度真实的模拟环境。PLT-D3的应用领域主要集中在自动驾驶技术的训练和测试，特别是在极端和动态天气条件下的性能评估。

The PLT-D3 dataset was created by Pilotier Toronto, focusing on enhancing the performance of autonomous driving systems under diverse weather conditions. It contains high-fidelity stereo depth and ground-truth scene flow data, generated using Unreal Engine 5, and covers scenarios with rain, snow, fog as well as various lighting conditions. The dataset includes synchronized high-resolution stereo image sequences, aiming to address the shortcomings of existing datasets in complex weather simulation. During the development process, the robust functionalities of UE5 were utilized to build highly realistic simulated environments. The primary application fields of PLT-D3 are centered on the training and testing of autonomous driving technologies, particularly performance evaluation under extreme and dynamic weather conditions.