Input data is finite, 2D, real and positive but iptcheckinput error

Input data is finite, 2D, real and positive but iptcheckinput error in the "edgetaper" and "deconvlucy" gives the following error messages:

??? Error using ==> iptcheckinput Function EDGETAPER expected its first input, I, to be finite.

Error in ==> edgetaper>parse_inputs at 117 iptcheckinput(I,{'uint8','uint16','int16','double','single'},{'finite', ...

Error in ==> edgetaper at 35 [I, PSF, sizeI, classI, sizePSF, numNSdim] = parse_inputs(varargin{:});

I know origdata is a 2D array of real doubles between 0 and 256 as I have done the following checks in the command line

>> sum(sum(abs(imag(origdata))))

ans =

     0

>> min(min(origdata))

ans =

     0
>> max(max(origdata))

ans =

   256
>> ndims(origdata)

ans =

     2

I am struggling to see any way to deconvolute this image (I have tried the 3 different deconvolution functions but they all throw up the same errors)

In the error messages above, the error appears for edgetaper, but if I remove this and try putting my data straight into any of the deconv functions, the same errors come up.

The code that throws up the error:

I = edgetaper(origdata,gausspsf); datadeconv = deconvlucy(I,gausspsf);

Answer 1

[Walter Roberson]

You should check

any(isnan(origdata(:)))

Remember that max() and min() ignore NaN