遇见数据集

Base images for 360 degrees panorama with camera poses

收藏
Mendeley Data2024-03-27 更新2024-06-29 收录
官方服务:

资源简介:

This dataset presents base images for 360 degrees image construction through stitching. The HD resolution base images are already undistorted. This dataset is meant for 360 degrees panorama generation and optimization techniques comparison. Each dataset folder presents a "cameras.sfm" file that contains the camera pose for every acquired image, which is estimated from IMU and servo motor sensors. The file presents the total number of images in the first line. Considering the rotation matrix as 'R' (with dimensions 3x3) , translation vector as 't' (with dimensions 3x1), 'fx' and 'fy' as the focal lengths in both X and Y axis, and 'cx' and 'cy' as the principal point coordinates in the image, every following line is written as: "image_name.png" R11 R12 R13 R21 R22 R23 R31 R32 R33 t11 t21 t31 fx fy cx cy with the numbers representing the position in the matrix and vector. In this way, every image has an approximated first guess for it's pose.

本数据集提供用于通过拼接构建360度图像的基础图像。该高清分辨率基础图像均已完成畸变校正,适用于360度全景生成及优化技术的对比研究。每个数据集文件夹均包含一个名为cameras.sfm的文件,其中存储了每幅采集图像的相机位姿,该位姿通过惯性测量单元(IMU)与伺服电机传感器估算得到。该文件的首行标注了图像总数量。设旋转矩阵为R(尺寸为3×3)、平移向量为t(尺寸为3×1),fx与fy分别为X轴与Y轴方向的焦距,cx与cy为图像的主点坐标,则后续每一行的格式为:"image_name.png" R11 R12 R13 R21 R22 R23 R31 R32 R33 t11 t21 t31 fx fy cx cy,其中各数字分别对应矩阵与向量中的元素位置。通过该格式,每幅图像均对应其位姿的近似初始估计值。

创建时间:
2024-01-23
二维码
社区交流群
二维码
科研交流群
商业服务