FEDORA
收藏资源简介:
FEDORA是由普渡大学创建的一个完全合成的飞行事件数据集,旨在支持自主飞行操作中的感知任务。该数据集包含来自帧基相机、事件基相机和惯性测量单元(IMU)的原始数据,以及深度、姿态和光流的高速率地面真实数据。FEDORA通过提供多种光照条件、不同运动和环境条件下的序列,旨在推动新型导航算法的研究。数据集的创建过程利用了开源物理引擎Gazebo进行模拟,确保了数据的真实性和多样性。FEDORA的应用领域主要集中在无人机导航,特别是解决高速运动和低延迟响应的需求。
FEDORA is a fully synthetic flight event dataset created by Purdue University, designed to support perception tasks in autonomous flight operations. This dataset contains raw data from frame-based cameras, event-based cameras, and inertial measurement unit (IMU), as well as high-rate ground truth data for depth, pose, and optical flow. By providing sequences under various lighting conditions, different motion and environmental scenarios, FEDORA aims to advance research on novel navigation algorithms. The dataset was developed using the open-source physics engine Gazebo for simulation, ensuring the authenticity and diversity of the data. The application fields of FEDORA mainly focus on unmanned aerial vehicle (UAV) navigation, particularly addressing the requirements of high-speed motion and low-latency response.

- 1FEDORA: Flying Event Dataset fOr Reactive behAvior普渡大学 · 2024年



