因为做卷积反投影,所以需要推导滤波函数在时域的表达。
H R L ( w ) = { ∣ w ∣ ∣ w ∣ ≤ w m 0 ∣ w ∣ > w m H_{RL}(w)=\left\{ \begin{array}{c} |w| && |w|\leq w_m\\ 0 && |w|>w_m \end{array} \right. HRL(w)={∣w∣0∣w∣≤wm∣w∣>wm
可以通过mathematical推导其Fourier逆变换。
语句如下:
最初做就使用R-L滤波器,离散化采样有:
h ( n d ) = ( c o s ( n π ) − 1 ) 2 π 2 ( n d ) 2 h(nd)=\frac{(cos(n\pi)-1)}{2\pi^2(nd)^2} h(nd)=2π2(nd)2(cos(nπ)−1)
分情况讨论:
频域形式为:
H S L ( w ) = ∣ w ∣ s i n c ( π w 2 w m ) r e c t ( w w m ) H_{SL}(w)=|w|sinc(\frac{\pi w}{2w_m})rect(\frac{w}{w_m}) HSL(w)=∣w∣sinc(2wmπw)rect(wmw)
其中:
r e c t ( x ) = { 1 ∣ x ∣ ≤ 1 0 o t h e r w i s e rect(x)=\left\{ \begin{array}{c} 1 && |x|\leq1\\ 0 &&otherwise \end{array} \right. rect(x)={10∣x∣≤1otherwise
频域形式为:
H C L ( w ) = ∣ w ∣ c o s ( π w 2 w m ) r e c t ( w w m ) H_{CL}(w)=|w|cos(\frac{\pi w}{2w_m})rect(\frac{w}{w_m}) HCL(w)=∣w∣cos(2wmπw)rect(wmw)
H C L ( w ) = ∣ w ∣ ( c o s ( π w 2 w m ) ( 1 − α ) + α ) r e c t ( w w m ) H_{CL}(w)=|w|(cos(\frac{\pi w}{2w_m})(1-\alpha)+\alpha)rect(\frac{w}{w_m}) HCL(w)=∣w∣(cos(2wmπw)(1−α)+α)rect(wmw)
CBP: f ( x 1 , x 2 ) = ∫ 0 π [ f ^ ( r , ϕ ) ∗ h ( r ) ] r = ( x 1 , x 2 ) ⋅ ϕ d ϕ f(x_1,x_2)=\int_0^{\pi}[\hat{f}(r,\phi)*h(r)]_{r=(x_1,x_2)\cdot\phi}d\phi f(x1,x2)=∫0π[f^(r,ϕ)∗h(r)]r=(x1,x2)⋅ϕdϕ
第一步时域滤波
g ( n d , ϕ ) = ∑ m = − ∞ ∞ f ^ ( n d , ϕ ) h ( n d − m d ) d g(nd,\phi)=\sum\limits_{m=-\infty}^{\infty}\hat{f}(nd,\phi)h(nd-md)d g(nd,ϕ)=m=−∞∑∞f^(nd,ϕ)h(nd−md)d
对其离散化有:
f ( x 1 , x 2 ) ≈ ∑ m = 1 M g ( r , ϕ m ) r = ( x 1 , x 2 ) ⋅ ϕ m Δ ϕ f(x_1,x_2)\approx \sum\limits_{m=1}^M g(r,\phi_m)_{r=(x_1,x_2)\cdot \phi_m}\Delta\phi f(x1,x2)≈m=1∑Mg(r,ϕm)r=(x1,x2)⋅ϕmΔϕ