The essential matrix relates two calibrated camera views. 本质矩阵用于关联两幅已标定相机视图。
After estimating the essential matrix from matched keypoints, we decomposed it to recover the relative rotation and translation direction between the cameras. 在用匹配到的关键点估计出本质矩阵之后,我们将其分解以恢复相机之间的相对旋转以及平移方向。