二维高斯正态分布函数(转)

二维高斯正态分布函数(原创)

二维高斯正态分布函数在很多地方都用的到，比如说在滤波中，自己编了个，但感觉IDL中应该有现成的函数？？（我没找到）。如有，请高手指点。

;-----------------------------

;二维高斯正态分布函数

;x,y:输入的一维数组(维数应该相同)

;x0,y0:均值

;sd:标准差

function Gaussian2DDistribution,x,y,x0,y0,sd

xsize=N_ELEMENTS(x)

ysize=N_ELEMENTS(y)

if (xsize ne ysize) then begin

Result = DIALOG_MESSAGE('输入的X和Y数组的维数不相同！',/error,title='警告')

return,0

endif

pz=fltarr(xsize,ysize)

for i=0,xsize-1 do begin

for j=0,ysize-1 do begin

a=-((x[i]-x0)*(x[i]-x0)+(y[j]-y0)*(y[j]-y0))/(2.0*sd*sd)

pz[i,j]=exp(a)/(2*3.1415926*sd*sd)

endfor

endfor

return,pz

end

;-------------------------------

pro Gaussian2DDistribution_test

x=FINDGEN(11)-5.0

y=FINDGEN(11)-5.0

z=Gaussian2DDistribution(x,y,0.0,0.0,2.5)

WINDOW,/free, XSIZE=500, YSIZE=500,title='Gaussian 2D Distribution'

SHADE_SURF, z ,x,y

SURFACE, Z,x,y, XST = 4, YST = 4, ZST = 4, /NOERASE

print,total(z)

end

-------------------------------------------------

补充：后来找了找，在CANNY函数中找到了相同的内容，把它贴出来吧

;;---------------------------------------------------------------------------

;; Canny_gaussian

;;

;; Purpose:

;; Create a guassian kernal for smoothing

;;

;; Parameters:

;; SIGMA - sigma value

;;

;; WIDTH - desired width of the kernel

;;

;; Keywords:

;; NONE

;;

FUNCTION canny_gaussian, sigma, width

IF n_elements(sigma) NE 1 THEN sigma = 1.0d

;; if not specified create a 5x5 kernel

IF (n_elements(width) EQ 0) THEN width = 5

;; create X and Y indices

x = (dindgen(width)-width/2) # replicate(1, width)

y = transpose(x)

;; create kernel

kernel = exp(-((x^2)+(y^2))/(2*double(sigma)^2)) / (sqrt(2.0*!pi) * double(sigma))

;; give it an appropriate scale factor

scale = 1/min(kernel) < 16384

kernel = TEMPORARY(kernel)*scale + 1d-6

return, long(kernel)

END