Hi every one I am working on Wiener deconvolution. my aim is to deconvolve the refelectivity function. I am using threshold method in inverse filtering for the deconvolution. Hf = H.*(abs(H)>0)+1/gamma*(abs(H)==0); iHf = 1./Hf; iHf = iHf.*(abs(H)*gamma>1)+gamma*abs(Hf).*iHf.*(abs(Hf)*gamma<=1); W = iHf.*Yyy./(Yyy+ sigma); %Inverse and denoise filters in cascade the code is working when gamma = 1 and signal =0; But when I change the parameters gamma there is no deconvolved output. can anybody help me in estmating the parameter gamma. or please mention refference to find out the gamma;

# wiener deconvolution

Started by ●November 3, 2009

Reply by ●November 3, 20092009-11-03

On 3 Nov, 19:15, "ramjanagam" <ramamoorthy...@yahoo.com> wrote:> Hi every one > > I am working on Wiener deconvolution. my aim is to deconvolve the > refelectivity function. I am using threshold method in inverse filtering > for the deconvolution. > > Hf = H.*(abs(H)>0)+1/gamma*(abs(H)==0); > iHf = 1./Hf; > iHf = iHf.*(abs(H)*gamma>1)+gamma*abs(Hf).*iHf.*(abs(Hf)*gamma<=1); > > W = iHf.*Yyy./(Yyy+ sigma); %Inverse and denoise filters in cascade > > the code is working when gamma = 1 and signal =0; > > But when I change the parameters gamma there is no deconvolved output.If the algorithm breaks down when you change one parameter, it is likely an implementation blunder somewhere. First of all make sure your impleentation is correct with respect to the algorithm you have chosen. Only after you have verified this, does it make sense to ask about optimum choises of parameters. And yes, verify your implementation with simulated data where the filter is known to work - preferably data where analytic solutions are known. Rune