Synthetic satellite point cloud dataset_part3
Description
Part3 of Synthetic satellite point cloud dataset. The dataset was developed by National Key Laboratory of Aerospace Flight Dynamics (AFDL) at Northwestern Polytechnical University (NWPU). This is a large-scale synthetic spacecraft point cloud dataset, comprising 256 satellite models and more than 139,000 point clouds with accurate pose data. The dataset is intended for training neural networks for point cloud registration. 该数据集是合成卫星数据集的第3部分。数据集由西北工业大学航天飞行动力学实验室开发。该数据集包含了256个卫星模型、超过139000个点云，每个点云带有位姿数据。该点云数据集用于点云配准相关的神经网络训练。
Files
Steps to reproduce
Please sequentially extract the three zip files to obtain 256 synthetic satellite folders. Each folder contains an STL 3D model, a TXT pose file, and PCD point cloud files. In the pose file, each line corresponds to one PCD file and consists of the timestamp, Euler angles in the RPY order, and the x/y/z position. Note that the transformation matrix from the world coordinate frame to the camera coordinate frame is[ [0.0, 0.0, -1.0, 0.0], [-1.0, 0.0, 0.0, 0.0], [0.0, 1.0, 0.0, 0.0], [0.0, 0.0, 0.0, 1.0] ]. During training, please take care to select dense point clouds and samples with relatively small pose differences; otherwise, it will be difficult for the network to converge. 请依次解压缩三个zip文件，从而获得256个合成卫星文件夹。每个文件夹下包含stl格式的3D模型、txt格式的位姿文件，以及pcd点云文件。其中位姿文件的每一行数据对应一个pcd文件，分别为时间、rpy通道的欧拉角、x/y/z数据。需要注意的是，世界坐标系到相机坐标系的变换矩阵是[ [0.0, 0.0, -1.0, 0.0], [-1.0, 0.0, 0.0, 0.0], [0.0, 1.0, 0.0, 0.0], [0.0, 0.0, 0.0, 1.0] ]