NAME: FFT_CONVOLUTION PURPOSE: Computes the convolution of two or many 1D to 3D arrays by using the FFT. DESCRIPTION: This function computes in one time the convolution of an array with one or more convolution kernels, e.g. F*K1*K2*...Kn. 1 dimensional to 3 dimensional arrays are supported. CATEGORY: Mathematics. CALLING SEQUENCE: Result = FFT_CONVOLUTION(array, kernel, /DOUBLE) INPUTS: Array: 1-D to 3-D array to be convolved. Kernel: 1-D to 3-D array to be used as convolution kernel. Kernel Must have the SAME SIZE as Array. For multiple convolutions Kernel must have the dimension of Array + 1 and must have one convolution kernel for each element of the Last dimension: e.g.: for F(x,y)*K1(x,y)*K2(x,y) where F is of LX*LY size the kernel is an array of [LX, LY, 2] in size. The Kernel is always centered on the Array points to make convolution. KEYWORD PARAMETERS: DOUBLE: if set uses the double precision for the FFT. MODIFICATION HISTORY: Feb 2004 - Gianluca Li Causi & Massimo De Luca, INAF - Rome Astronomical Observatory licausi@mporzio.astro.it http://www.mporzio.astro.it/~licausi/