Hydra Benchmark Dataset
收藏资源简介:
Hydra基准数据集是用于手眼校准的无标记RGB-D成像方法的基准数据集。数据集包含三个串联操作机和两个RGB-D相机的综合实验结果,展示了该方法在机器人校准领域的应用。数据集共包含270个样本,通过随机选择三个机器人配置,实现了约90%的成功校准,并展示了比基于标记和无标记基线更高的收敛速度。此外,数据集还报告了比其他无标记方法快两个数量级的收敛时间。数据集和代码已开源,并提供了ROS 2集成,以促进部署。
The Hydra benchmark dataset is a benchmark dataset for marker-free RGB-D imaging methods used for hand-eye calibration. The dataset contains comprehensive experimental results from three serial manipulators and two RGB-D cameras, demonstrating the application of the proposed method in the field of robotic calibration. It includes a total of 270 samples, where randomly selecting three robot configurations per sample achieves an approximate 90% successful calibration rate, and exhibits a higher convergence speed than both marker-based and marker-free baseline methods. Additionally, the dataset shows that the convergence time is two orders of magnitude faster than that of other existing marker-free methods. Furthermore, the dataset and accompanying code are open-sourced, with ROS 2 integration provided to facilitate practical deployment.




