A Comprehensive View-Dependent Dataset for Objects Reconstruction - Kinect Azure Set

The dataset was collected by moving the Kinect Azure camera above a set of real objects located on the ArUco board that was used to estimate the camera pose. To estimate the camera poses for the objects from the subset of real objects, we used the ArUco board, which consisted of a grid of 7x5 unique ArUco markers. We placed the object on the ArUco board and utilized the OpenCV library methods to estimate the camera poses located around the object and the marker. For estimating camera poses and creating the 3D model of the objects, we employed an ArUco board featuring a grid of 7x5 distinctive ArUco markers. By situating the object on the ArUco board, we utilized OpenCV library methods to calculate the camera poses positioned both around the object and the markers.

本数据集通过将Kinect Azure相机（Kinect Azure）置于放置于ArUco板（ArUco board）上的一组真实物体上方并移动采集得到，该ArUco板用于估算相机位姿。为估算该真实物体子集内各物体对应的相机位姿，我们采用了由7×5个唯一ArUco标记（ArUco marker）组成的ArUco板：将物体放置于该ArUco板上后，借助OpenCV库（OpenCV）的相关方法即可估算物体与标记周围的相机位姿。在相机位姿估算与物体三维模型构建的任务中，我们同样使用了带有7×5个独特ArUco标记网格的ArUco板，并通过将物体置于该板上，利用OpenCV库的方法计算得到分布于物体与标记周围的相机位姿。