next up previous contents
Next: B.2 Rectangular Patterns Up: B. Analytical Fourier Transformations Previous: B. Analytical Fourier Transformations

  
B.1 Triangular Patterns

An arbitrary triangle consisting of three points {(x1, y1),(x2, y2),(x3, y3)} in $ \mathbb{R}$2 can always be transformed to the unit triangle {(0, 0),(1, 0),(0, 1)} by

 
$\displaystyle \begin{pmatrix}x \\ y \end{pmatrix} = \underline{\mathbf{A}} \beg...
...thbf{A}} = \begin{pmatrix}x_2-x_1 & x_3-x_1 \\ y_2-y_1 & y_3-y_1 \end{pmatrix}.$ (B.2)

The module of the determinant of the transformation matrix $ \underline{\mathbf{A}}$ equals

$\displaystyle \vert\det\underline{\mathbf{A}}\,\vert = \vert(x_2-x_1)(y_3-y_1) - (y_2-y_1)(x_3-x_1)\vert.$ (B.3)

Using (B.2), the integral (B.1) writes for triangular patterns to

$\displaystyle \begin{aligned}\mathcal{T}_{nm} &= \frac{1}{ab} \iint\limits_{(x,...
... d\eta} _{\displaystyle I_{\mathcal{T}}(\alpha_{nm},\beta_{nm})}, \end{aligned}$    

whereby TRI-1 refers to the unit triangle and the coefficients $ \alpha_{nm}^{}$, $ \beta_{nm}^{}$ and $ \gamma_{nm}^{}$ are given by

$\displaystyle \renewcommand{\arraystretch}{2}\alpha_{nm}$ $\displaystyle = n(x_2-x_1)/a + m(y_2-y_1)/b$    
$\displaystyle \beta_{nm}$ $\displaystyle = n(x_3-x_1)/a + m(y_3-y_1)/b$ (B.4)
$\displaystyle \gamma_{nm}$ $\displaystyle = nx_1/a + my_1/b.$    

The function I$\scriptstyle \mathcal {T}$($ \alpha_{nm}^{}$,$ \beta_{nm}^{}$) is calculated as follows:a

 
$\displaystyle I_{\mathcal{T}}(\alpha_{nm},\beta_{nm}) = \begin{cases}\displayst...
...a_{nm}-\alpha_{nm}} e^{-j\pi\beta_{nm}} \right] & \text{otherwise.} \end{cases}$ (B.5)



Footnotes

... follows:a
The function si(x) is defined as si(x)$ \overset{\mathrm{def}}{=} \sin x/x$.

next up previous contents
Next: B.2 Rectangular Patterns Up: B. Analytical Fourier Transformations Previous: B. Analytical Fourier Transformations
Heinrich Kirchauer, Institute for Microelectronics, TU Vienna
1998-04-17