Creating a 3D surface plot from array data

I have this script

clear
img = imread('cameraman.tif');
Mean = 0.01:0.01:0.05;
hsize = 3:1:25;

for m = 1:length(Mean)
    for h = 1:length(hsize)

V=0.01;
noise = imnoise(img,'gaussian',Mean(m),V);
[peaksnr, snr] = psnr(noise,img);
nsr = 1/snr;

PSF = fspecial('average',hsize(h));
AverageBlur = imfilter(img,PSF,'replicate','same','conv');    


NoiseyBlurred = imfilter(noise,PSF,'replicate','same','conv');
NoiseyBlurredImage = NoiseyBlurred;
BlindDeconv = deconvblind(NoiseyBlurredImage,PSF,15);
LucyFilter = deconvlucy(NoiseyBlurredImage,PSF,15);
WNRFilter = deconvwnr(NoiseyBlurredImage,PSF,nsr);
RegularFilter = deconvreg(NoiseyBlurredImage,PSF,NPowerInt);

PSNRBlurred = psnr(NoiseyBlurredImage,img);
PSNRBlind = psnr(BlindDeconv,img);
PSNRLucy = psnr(LucyFilter,img);
PSNRWNR = psnr(WNRFilter,img);
PSNRReg = psnr(RegularFilter,img);

Blur(h,m)= (40-PSNRBlurred)/40;
Blind(h,m)=PSNRBlind;
Lucy(h,m)=PSNRLucy;
WNR(h,m)=PSNRWNR;
Reg(h,m)=PSNRReg;

    end
end

It saves the PSNR value for each hsize and mean in 25x5 array like this:

Lucy =

21.6467 21.4919 21.1253 20.7971 20.3298

22.2624 22.1446 21.7016 21.3296 20.7955

22.4250 22.1990 21.8411 21.4450 21.0195

Each column has 25 values which i cut short for space. The blurred error (denoted 'Blur') saves in the same way.

I am trying to plot a 3D surface/mesh of the data, where for each column PSNR vs error is plotted, rather than just a 2D line plot with 5 lines over the top of eachother. Is there a way to do this?

Edit: Attached Image of 2D plot to help with clarity of the question. Can i make a graph like this but 3D to be able to see all the lines (ideally as a surface) and with the new axis being the 'Mean'. The 5 lines are due to the 5 Mean values. The error and psnr value changes with each hsize, i.e. 23 x,y values per line. Allowing me to see how both hsize and mean affect the PSNR value.

Apologies for the lack of clarity and specificity in the original question, hopefully this helps clear things up.

Answer 1

[Star Strider]

Try this:

figure
surf(Lucy)
grid on

or:

figure
mesh(Lucy)
grid on

The ‘x’ and ‘y’ axes are in units of the relevant indices, since ‘x’ and ‘y’ vectors or matrices defining them otherwise are not provided.

[Steven Harkness]

Thanks for the quick reply!

I just tried it and although it is the right kind of plot I also need to include the values from Blur in the plot too. I looked into surf and tried surf(Lucy,Blur,z) with z = 1:1:5 to represent the 5 columns but that wouldnt plot as z was a vector not a matrix.

[Star Strider]

My pleasure!

You need to be more specific. If you want to plot both ‘Lucy’ and ‘Blur’, on the same axes, probably the easiest way is:

figure
surf(Lucy)
hold on
surf(Blur)
hold off
grid on

.

[Steven Harkness]

Apologies for the question not being specifc and clear enough. I've edited it to try and help clear things up. That did plot them both but not in the way i meant (again sorry for not making it clear), the edit in the question should show what im trying to do. Thanks again

[Star Strider]

The ribbon function may be what you want. Since you mentioned that your arrays are (25x5), this will plot them using the row size as the independent variable:

Lucy = [   21.6467   21.4919   21.1253   20.7971   20.3298
   22.2624   22.1446   21.7016   21.3296   20.7955
   22.4250   22.1990   21.8411   21.4450   21.0195];
figure
ribbon((1:size(Lucy,1)),Lucy, 0.1)
grid on

Another option:

x = (ones(size(Lucy,2),1)*(1:size(Lucy,1)))';
y = ones(size(Lucy,1),1)*(1:size(Lucy,2));
x = [x; nan(1,size(x,2))];
y = [y; nan(1,size(y,2))];
Lucy = [Lucy; nan(1,size(Lucy,2))];
figure
plot3(x(:),y(:),Lucy(:))
grid on

Experiment with them to get the result you want.

.

[Steven Harkness]

Ribbon is perfect, thank you!

[Star Strider]

As always, my pleasure!