《视觉SLAM十四讲》学习笔记-第四讲部分习题的证明思路

本文链接：https://blog.csdn.net/luohuiwu/article/details/80719357

1 验证SO(3)、SE(3) 和Sim(3) 关于乘法成群
证明：
先看SO(3). 定义为：

S O (3) = {R \in R 3 \times 3 | R R ⊤ = I, det (R) = 1}

$SO(3) = \{\mathbf{R}\in \mathbb{R} ^{3\times3} | \mathbf{R}\mathbf{R}^\top = \mathbf{I}, \det(\mathbf{R})=1 \}$

假设 $\mathbf{R}_1, \mathbf{R}_2\in SO(3)$ , 先证明 $\mathbf{R}_1*\mathbf{R}_2\in SO(3)$ :

$R 1 * R 2 * (R 1 * R 2) ⊤ = R 1 * R 2 R ⊤ 2 * R ⊤ 1 = I$ $\mathbf{R}_1*\mathbf{R}_2*(\mathbf{R}_1*\mathbf{R}_2)^\top = \mathbf{R}_1*\mathbf{R}_2\mathbf{R}_2^\top*\mathbf{R}_1^\top = \mathbf{I}$
再证其行列式为1:
$det (R 1 * R 2) = det (R 1) \cdot det (R 2) = 1$ $\det(\mathbf{R}_1*\mathbf{R}_2) = \det(\mathbf{R}_1) \cdot \det(\mathbf{R}_2) = 1$
所以 $\mathbf{R}_1, \mathbf{R}_2\in SO(3) \Rightarrow \mathbf{R}_1*\mathbf{R}_2\in SO(3)$ .
根据矩阵乘法，显然：

$(R 1 * R 2) * R 3 = R 1 * (R 2 * R 3)$ $(\mathbf{R}_1 * \mathbf{R}_2) * \mathbf{R}_3 = \mathbf{R}_1 * (\mathbf{R}_2 * \mathbf{R}_3 )$
存在 $\mathbf{I}_0 \in \mathbb{R}^{3\times 3}$ (单位矩阵), 使得:

$I 0 * R = R * I 0 = R$ $\mathbf{I}_0 * \mathbf{R} = \mathbf{R} * \mathbf{I}_0 = \mathbf{R}$
所以玄元为 $\mathbf{I}_0$ .
根据定义 $\mathbf{R} \mathbf{R}^\top = \mathbf{I}$ , 所以逆为

$R - 1 = R ⊤ \in S O (3)$ $\mathbf{R}^{-1} = \mathbf{R}^\top \in SO(3)$

SE(3)的定义为：

S E (3) = {T = [R 0 ⃗ t ⃗ 1] \in R 4 \times 4 | R \in S O (3), t ⃗ \in R 3}

$SE(3) = \{\mathbf{T} = \begin{bmatrix} \mathbf{R} & \vec{t}\\ \vec{0} & 1 \end{bmatrix} \in \mathbb{R}^{4\times 4} | \mathbf{R} \in SO(3), \vec{t}\in \mathbb{R}^3 \}$
- 假设

T1,T2∈SE(3) T 1 , T 2 ∈ S E ( 3 ) $\mathbf{T}_1, \mathbf{T}_2\in SE(3)$ , 证明

T1∗T2∈SE(3) T 1 ∗ T 2 ∈ S E ( 3 ) $\mathbf{T}_1*\mathbf{T}_2\in SE(3)$ :

T 1 * T 2 = [R 1 0 ⃗ t ⃗ 1 1] * [R 2 0 ⃗ t ⃗ 2 1] = [R 1 * R 2 0 ⃗ R 1 t ⃗ 2 + t ⃗ 1 1]

$\begin{aligned} \mathbf{T}_1*\mathbf{T}_2 &= \begin{bmatrix} \mathbf{R}_1 & \vec{t}_1\\ \vec{0} & 1 \end{bmatrix}* \begin{bmatrix} \mathbf{R}_2 & \vec{t}_2\\ \vec{0} & 1 \end{bmatrix}\\ &=\begin{bmatrix} \mathbf{R}_1*\mathbf{R}_2 & \mathbf{R}_1\vec{t}_2+\vec{t}_1\\ \vec{0} & 1 \end{bmatrix} \end{aligned}$
上述式子中，

R1,R2∈SO(3) R 1 , R 2 ∈ S O ( 3 ) $\mathbf{R}_1, \mathbf{R}_2\in SO(3)$ ,

R1∗R2∈SO(3) R 1 ∗ R 2 ∈ S O ( 3 ) $\mathbf{R}_1*\mathbf{R}_2\in SO(3)$ 已获得证明；

R1t⃗ 2+t⃗ 1∈R3 R 1 t → 2 + t → 1 ∈ R 3 $\mathbf{R}_1\vec{t}_2+\vec{t}_1 \in \mathbb{R}^3$ .
满足SE(3)的定义。

同样根据矩阵乘法原理， $\mathbf{R}_1, \mathbf{R}_2, \mathbf{R}_3\in SO(3)$ ,
$(R 1 * R 2) * R 3 = R 1 * (R 2 * R 3)$ $(\mathbf{R}_1 * \mathbf{R}_2) * \mathbf{R}_3 = \mathbf{R}_1 * (\mathbf{R}_2 * \mathbf{R}_3)$
SE(3)的玄元为 $\mathbf{I}_0 \in \mathbb{R}^{4\times 4}$
SE(3)在乘法下的逆为：
$T - 1 = [R 0 ⃗ t ⃗ 1] - 1 = [R ⊤ 0 ⃗ - R - 1 t ⃗ 1] = [R ⊤ 0 ⃗ - R ⊤ t ⃗ 1]$ $\mathbf{T}^{-1}= \begin{bmatrix} \mathbf{R} & \vec{t} \\ \vec{0} & 1 \end{bmatrix}^{-1}= \begin{bmatrix} \mathbf{R}^\top & -\mathbf{R}^{-1}\vec{t}\\ \vec{0} & 1 \end{bmatrix}= \begin{bmatrix} \mathbf{R}^\top & -\mathbf{R}^\top\vec{t}\\ \vec{0} & 1 \end{bmatrix}$
Sim(3)的证明与SE(3)很类似，限于篇幅就不展开了。

2 验证 $(\mathbb{R}^3, \mathbb{R}, \times)$ 构成李代数
前面文章已证。

3 **验证 $\mathfrak{so}(3)$ 和 $\mathfrak{se}(3)$ 满足李代数要求的性质
$\mathfrak{so}(3)$ 已在前面文章中证明。下面只证明 $\mathfrak{se}(3)$ **.

4 证明式(4.20)与(4.21).
- 证明式(4.20) ：

a ⃗ \land = ⎛ ⎝ ⎜ 0 a 3 - a 2 - a 3 0 a 1 a 2 - a 1 0 ⎞ ⎠ ⎟

$\vec{a}^\wedge = \left(\begin{array}{ccc} 0 & -a_{3} & a_{2}\\ a_{3} & 0 & -a_{1}\\ -a_{2} & a_{1} & 0 \end{array}\right)$
所以，

a ⃗ \land a ⃗ \land = ⎛ ⎝ ⎜ - a 2 2 - a 3 2 a 1 a 2 a 1 a 3 a 1 a 2 - a 1 2 - a 3 2 a 2 a 3 a 1 a 3 a 2 a 3 - a 1 2 - a 2 2 ⎞ ⎠ ⎟

$\vec{a}^\wedge\vec{a}^\wedge=\left(\begin{array}{ccc} -{a_{2}}^2-{a_{3}}^2 & a_{1}\,a_{2} & a_{1}\,a_{3}\\ a_{1}\,a_{2} & -{a_{1}}^2-{a_{3}}^2 & a_{2}\,a_{3}\\ a_{1}\,a_{3} & a_{2}\,a_{3} & -{a_{1}}^2-{a_{2}}^2 \end{array}\right)$
另外，

a ⃗ a ⃗ ⊤ - I = ⎛ ⎝ ⎜ a 1 c o n j (a 1) - 1 a 2 c o n j (a 1) a 3 c o n j (a 1) a 1 c o n j (a 2) a 2 c o n j (a 2) - 1 a 3 c o n j (a 2) a 1 c o n j (a 3) a 2 c o n j (a 3) a 3 c o n j (a 3) - 1 ⎞ ⎠ ⎟

$\vec{a}\vec{a}^\top-\mathbf{I}=\left(\begin{array}{ccc} a_{1}\,\mathrm{conj}\left(a_{1}\right)-1 & a_{1}\,\mathrm{conj}\left(a_{2}\right) & a_{1}\,\mathrm{conj}\left(a_{3}\right)\\ a_{2}\,\mathrm{conj}\left(a_{1}\right) & a_{2}\,\mathrm{conj}\left(a_{2}\right)-1 & a_{2}\,\mathrm{conj}\left(a_{3}\right)\\ a_{3}\,\mathrm{conj}\left(a_{1}\right) & a_{3}\,\mathrm{conj}\left(a_{2}\right) & a_{3}\,\mathrm{conj}\left(a_{3}\right)-1 \end{array}\right)$
因为

a⃗ a → $\vec{a}$ 单位长为1，所以

a21+a22+a23=1 a 1 2 + a 2 2 + a 3 2 = 1 $a_1^2 + a_2^2 + a_3^2 = 1$ ，所以

a⃗ ∧a⃗ ∧=a⃗ a⃗ ⊤−I a → ∧ a → ∧ = a → a → ⊤ − I $\vec{a}^\wedge\vec{a}^\wedge = \vec{a}\vec{a}^\top-\mathbf{I}$ .

证明式(4.21)
$a ⃗ \land a ⃗ \land a ⃗ \land = ⎛ ⎝ ⎜ ⎜ 0 - a 3 a 2 2 - a 3 (a 1 2 + a 3 2) a 2 a 3 2 + a 2 (a 1 2 + a 2 2) a 3 a 1 2 + a 3 (a 2 2 + a 3 2) 0 - a 1 a 3 2 - a 1 (a 1 2 + a 2 2) - a 2 a 1 2 - a 2 (a 2 2 + a 3 2) a 1 a 2 2 + a 1 (a 1 2 + a 3 2) 0 ⎞ ⎠ ⎟ ⎟$ $\vec{a}^\wedge\vec{a}^\wedge\vec{a}^\wedge=\left(\begin{array}{ccc} 0 & a_{3}\,{a_{1}}^2+a_{3}\,\left({a_{2}}^2+{a_{3}}^2\right) & -a_{2}\,{a_{1}}^2-a_{2}\,\left({a_{2}}^2+{a_{3}}^2\right)\\ -a_{3}\,{a_{2}}^2-a_{3}\,\left({a_{1}}^2+{a_{3}}^2\right) & 0 & a_{1}\,{a_{2}}^2+a_{1}\,\left({a_{1}}^2+{a_{3}}^2\right)\\ a_{2}\,{a_{3}}^2+a_{2}\,\left({a_{1}}^2+{a_{2}}^2\right) & -a_{1}\,{a_{3}}^2-a_{1}\,\left({a_{1}}^2+{a_{2}}^2\right) & 0 \end{array}\right)$
注意到 $a_1^2 + a_2^2 + a_3^2 = 1$ ，且
$- a ⃗ \land = ⎛ ⎝ ⎜ 0 - a 3 a 2 a 3 0 - a 1 - a 2 a 1 0 ⎞ ⎠ ⎟$ $-\vec{a}^\wedge = \left(\begin{array}{ccc} 0 & a_{3} & -a_{2}\\ -a_{3} & 0 & a_{1}\\ a_{2} & -a_{1} & 0 \end{array}\right)$
所以: $\vec{a}^\wedge\vec{a}^\wedge\vec{a}^\wedge= -\vec{a}^\wedge$