Precise Synthetic Image and LiDAR (PreSIL) Dataset
收藏资源简介:
PreSIL数据集是由滑铁卢大学创建的,用于自动驾驶感知研究的高精度合成图像和激光雷达(LiDAR)数据集。该数据集包含超过50,000帧的高清图像,每帧图像均附带全分辨率深度信息、语义分割(图像)、点级分割(点云)以及详细的车辆和人员标注。数据收集过程完全自动化,无需人工标注。PreSIL数据集特别适用于3D物体检测,能够有效提升自动驾驶系统在复杂环境中的感知能力,解决自动驾驶中的动态物体检测问题。
The PreSIL dataset is a high-precision synthetic image and LiDAR dataset developed by the University of Waterloo for autonomous driving perception research. It includes over 50,000 frames of high-definition images, each accompanied by full-resolution depth maps, image-level semantic segmentation annotations, point-level segmentation annotations on point clouds, and detailed annotations for vehicles and pedestrians. The entire data collection process is fully automated, with no manual annotation required. The PreSIL dataset is specifically designed for 3D object detection, and can effectively enhance the perception performance of autonomous driving systems in complex environments, addressing the dynamic object detection challenge in autonomous driving.

- 1Precise Synthetic Image and LiDAR (PreSIL) Dataset for Autonomous Vehicle Perception滑铁卢大学 · 2019年



