2 预备知识 Preliminaries
In this chapter we give a condensed summary of the relevant mathematical concepts and notation.
在本章中,我们会简明扼要的给出(这篇论文)相关的数学定义和符号表示。
In particular, we summarize the representation of 3D poses as elements of Lie-Algebras (Sec. 2.1),
尤其在(2.1小节),我们会总结一下,三维空间位姿(如何)用李代数( Lie-Algebras)的方式表达。
derive direct image alignment as weighted least-squares minimization on Lie-manifolds (Sec. 2.2),
然后(2.2小节)推导出图像直接配准法的实质,即:李群—流体流形(Lie-manifolds)上的加权最小二乘的最小化(weighted least-squares minimization)(优化问题)
and briefly introduce propagation of uncertainty (Sec. 2.3).
并在(2.3小节)扼要说明不确定性(协方差矩阵或叫信息矩阵,译者额外添加备注)是如何传播的。(感谢范帝楷同学指点)
数学符号定义 Notation.
We denote matrices by bold, capital letters and vectors as bold, lower case letters ( ). (note that: Katex doesn't support bold greek letters)
“矩阵” 用粗体,大写字母表示,(比如:), ”向量” 用粗体, 小写字母表示(比如: )。
The n’th row of a matrix is denoted by .
矩阵的第n行用来表示。(这个表达形式在3.5小节的公式18. 19中出现,译者额外添加备注)
Images
the per-pixel inverse depth map
and the inverse depth variance map are written as functions,
where is the set of normalized pixel coordinates, i.e., they include the intrinsic camera calibration.
论文中提及的若干概念:
- 图像 (Image),
- (像素级别的)逆深度图 (简称逆深度图,per-pixel inverse depth map),
- 逆深度方差 (inverse depth variance map),
我们用数学映射关系的函数表达如下:
- 图像 ,( 从图像向一个实数的映射)
- 逆深度图 ( 逆深度图到正实数的映射)
- 逆深度方差 ( 逆深度方差到正实数的映射)
其中: , 是归一化(normalized)的像素(二维)坐标点集合,即:包含相机内参(intrinsic)其考虑了相机内参(intrinsic)。(译者:谢谢蔡育展老师指点)
Throughout the paper we use to denote the inverse of the depth of a point, i.e., .
在整篇论文中,三维空间中某一点的深度表示为 , 它的逆深度用 表示,两者的关系即: 。