Skip to content

Commit

Permalink
继续补充推导斜率积分#145 #156
Browse files Browse the repository at this point in the history
  • Loading branch information
kanition committed Mar 17, 2024
1 parent 514eddc commit 81fb095
Showing 1 changed file with 96 additions and 3 deletions.
99 changes: 96 additions & 3 deletions content/chap08ex01.tex
Original file line number Diff line number Diff line change
Expand Up @@ -568,7 +568,7 @@ \subsubsection*{Smith微面}
D({\bm\omega}_{\mathrm{h}})\mathrm{d}{\bm\omega}_{\mathrm{h}}\, .
\end{align}
于是
\begin{align}
\begin{align}\label{eq:08-ex01-g1_distance}
G_1^{\mathrm{d}}({\bm\omega}_{\mathrm{o}})
=\frac{\cos\theta_{\mathrm{o}}}
{\displaystyle\int\limits_{\varOmega}\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
Expand All @@ -591,7 +591,9 @@ \subsubsection*{Smith微面}
{\bm\omega}_{\mathrm{h}}=(x_{\mathrm{h}},y_{\mathrm{h}},z_{\mathrm{h}})
=(\sin\theta\cos\varphi,\sin\theta\sin\varphi,\cos\theta)\, ,
\end{align}
则该处附近的面元可以近似为以下平面
其中$\theta$为天顶角(即${\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{g}}=\cos\theta$),
$\varphi$为方位角
\sidenote{参见\reffig{5.ex01}。};则该处附近的面元可以近似为以下平面
\begin{align}
x_{\mathrm{h}}x+y_{\mathrm{h}}y+z_{\mathrm{h}}z=C\, .
\end{align}
Expand All @@ -611,7 +613,9 @@ \subsubsection*{Smith微面}
\sidenote{原文slope,这里作者想表达的是$z$$x$$y$的偏导数,
它和二维直角坐标系下直线斜率的形式很像,所以借用这个称呼。}。
反之,也可以根据斜率求出相应法线为
\begin{align}
\sidenote{注意到${\bm s}({\bm\omega}_{\mathrm{h}})={\bm s}(-{\bm\omega}_{\mathrm{h}})$
所以这里也可以是反向的结果,我们只是取其中一个。}
\begin{align}\label{eq:08-ex01-normals-by-slope}
{\bm\omega}_{\mathrm{h}}=\frac{1}{\sqrt{x_s^2+y_s^2+1}}(-x_s,-y_s,1)\, .
\end{align}

Expand Down Expand Up @@ -673,6 +677,95 @@ \subsubsection*{Smith微面}
=D({\bm\omega}_{\mathrm{h}})\cos^4\theta\, .
\end{align}

利用前面的结果,我们尝试将积分域从法线分布空间转化到斜率分布空间。
首先注意到,因为${\bm\omega}_{\mathrm{g}}=(0,0,1)$
所以结合\refeq{08-ex01-normals-by-slope}有
\begin{align}
{\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{g}}
=\frac{1}{\sqrt{x_s^2+y_s^2+1}}\, .
\end{align}
类似地,对于出射方向${\bm\omega}_{\mathrm{o}}=(x_{\mathrm{o}},y_{\mathrm{o}},z_{\mathrm{o}})$,则有
\begin{align}
\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
=\frac{(-x_{\mathrm{o}}x_s-y_{\mathrm{o}}y_s+z_{\mathrm{o}})
\chi(-x_{\mathrm{o}}x_s-y_{\mathrm{o}}y_s+z_{\mathrm{o}})}{\sqrt{x_s^2+y_s^2+1}}\, .
\end{align}
将上述两式与\refeq{08-ex01-P2D}结合可知
\begin{align}\label{eq:08-ex01-trans-normal-slope}
& \int\limits_{\varOmega}\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
D({\bm\omega}_{\mathrm{h}})\mathrm{d}{\bm\omega}_{\mathrm{h}}\nonumber \\
= & \int\limits_{\varOmega}\frac{\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)}
{\cos\theta}D({\bm\omega}_{\mathrm{h}})\cos\theta\mathrm{d}{\bm\omega}_{\mathrm{h}}\nonumber \\
= & \int_{-\infty}^{+\infty}\int_{-\infty}^{+\infty}
\frac{\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)}
{{\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{g}}}
P_{xy}(x_s,y_s)\mathrm{d}x_s\mathrm{d}y_s\nonumber \\
= & \int_{-\infty}^{+\infty}\int_{-\infty}^{+\infty}
(-x_sx_{\mathrm{o}}-y_sy_{\mathrm{o}}+z_{\mathrm{o}})
\chi(-x_sx_{\mathrm{o}}-y_sy_{\mathrm{o}}+z_{\mathrm{o}})
P_{xy}(x_s,y_s)\mathrm{d}x_s\mathrm{d}y_s\, .
\end{align}
为了简化后续推导,这里我们不妨设出射方向${\bm\omega}_{\mathrm{o}}$的方位角$\varphi=0$,于是
\begin{align}
{\bm\omega}_{\mathrm{o}}=(\sin\theta_{\mathrm{o}},0,\cos\theta_{\mathrm{o}})\, .
\end{align}
将其代入\refeq{08-ex01-trans-normal-slope}得
\begin{align}\label{eq:08-ex01-trans-1d-slope}
& \int\limits_{\varOmega}\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
D({\bm\omega}_{\mathrm{h}})\mathrm{d}{\bm\omega}_{\mathrm{h}}\nonumber \\
= & \int_{-\infty}^{+\infty}\int_{-\infty}^{+\infty}
(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})
\chi(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})
P_{xy}(x_s,y_s)\mathrm{d}x_s\mathrm{d}y_s\nonumber \\
= & \displaystyle\int_{-\infty}^{+\infty}
(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})
\chi(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})
\left(\int_{-\infty}^{+\infty}P_{xy}(x_s,y_s)\mathrm{d}y_s\right)\mathrm{d}x_s\nonumber \\
= & \displaystyle\int_{-\infty}^{+\infty}
(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})
\chi(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})P_x(x_s)\mathrm{d}x_s\, ,
\end{align}
其中
\begin{align}
P_x(x_s)=\int_{-\infty}^{+\infty}P_{xy}(x_s,y_s)\mathrm{d}y_s
\end{align}
是斜率沿着出射方向(这里假设方位角为零)的条件分布。
注意到\refeq{08-ex01-trans-1d-slope}中的示性函数把积分域限定在
\begin{align}
-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}}>0\, ,
\end{align}
也即
\begin{align}
x_s<\cot\theta_{\mathrm{o}}\, .
\end{align}
于是\refeq{08-ex01-trans-1d-slope}可进一步化简为
\begin{align}
\int\limits_{\varOmega}\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
D({\bm\omega}_{\mathrm{h}})\mathrm{d}{\bm\omega}_{\mathrm{h}}
=\int_{-\infty}^{\cot\theta_{\mathrm{o}}}
(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})P_x(x_s)\mathrm{d}x_s\, .
\end{align}
将上式代入\refeq{08-ex01-g1_distance},可得
\begin{align}
\cos\theta_{\mathrm{o}}
= & G_1^{\mathrm{d}}({\bm\omega}_{\mathrm{o}})
\int\limits_{\varOmega}\max({\bm\omega}_{\mathrm{h}}\cdot{\bm\omega}_{\mathrm{o}},0)
D({\bm\omega}_{\mathrm{h}})\mathrm{d}{\bm\omega}_{\mathrm{h}}\nonumber \\
= & G_1^{\mathrm{d}}({\bm\omega}_{\mathrm{o}})
\int_{-\infty}^{\cot\theta_{\mathrm{o}}}
(-x_s\sin\theta_{\mathrm{o}}+\cos\theta_{\mathrm{o}})P_x(x_s)\mathrm{d}x_s\, .
\end{align}
将上式两边同除以$\sin\theta_{\mathrm{o}}$,得到
\begin{align}
\cot\theta_{\mathrm{o}}=G_1^{\mathrm{d}}({\bm\omega}_{\mathrm{o}})
\int_{-\infty}^{\cot\theta_{\mathrm{o}}}
(-x_s+\cot\theta_{\mathrm{o}})P_x(x_s)\mathrm{d}x_s\, .
\end{align}
此外,我们假设微面法线的分布是中心对称的,即任意出射方向上的斜率均值都为零:
\begin{align}
\int_{-\infty}^{+\infty}x_sP_{x}(x_s)\mathrm{d}x_s=0\, ,
\end{align}

\subsection{典型微面分布函数的规范性证明}\label{sub:典型微面分布函数的规范性证明}
本节补充了\refeq{8.10}和\refeq{8.11}所给的
微面分布函数$D({\bm\omega}_{\mathrm{h}})$满足规范性要求的证明,即证明
Expand Down

0 comments on commit 81fb095

Please sign in to comment.