Resampling matrices in a loop using interp2

I'm trying to resample 200 matrices of 300*500 size to 30*50 using interp2. My objective is to load each resampled matrix into workspace after running the loop. In the code given here, it indicates there's an issue with how I use interp2 within the for-loop. Could someone please help me resolve this? Or any better approaches than this are also welcome.

-Thank you-

%filepath to directory where data is located

filepath='C:\Users\Task 1\';

%root part of filename (part that doesn't change each time)

filestem='VOD_Au_0dot1degree_';

for i=1:200

%loading files into a struct

A(i)=load(strcat(filepath,filestem,num2str(i),'.mat'));

%resampling

%new matrix column & row size

nc=50;nr=30;

%current matrix column & row size

c=500;r=300;

[C,R]=meshgrid(1:(c-1)/(nc-1):c,1:(r-1)/(nr-1):r);

downscaled_vod(i)=interp2(double(A(i).vod_monthly),C,R);

figure(i)=imagesc(downscaled_vod(i));

end

Answer 1

[Star Strider]

I am not certain what hte problem is, since the error messagd is not part of the provided information.

I suspect the problem is in saving a matrix to a scalar, defined by:

downscaled_vod(i)

That can be solved by saving it to a cell array element:

downscaled_vod{i}

The same sort of problem could be due to the ‘double(A(i).vod_monthly)’ reference, although without more information, I cannot determine that.

A = exp(-((1:300).'-150).^2/3000) * exp(-((1:500)-250).^2/5000);        % Create Matrix To Be Interpolated

figure
surf(A, 'EdgeColor','none')
colormap(turbo)
title('Original')

i = 1;

nc=50;
nr=30;
c=500;
r=300;
rv = linspace(1,r,nr);
cv = linspace(1,c,nc);
[C,R] = meshgrid(cv, rv);
downscaled.vod{i} = interp2(A, C, R);

figure
surf(downscaled.vod{i})
colormap(turbo)
title('Downscaled')

Using linspace to create the interpolation vectors is likely more efficien than the current code.

Also, put all these:

nc=50;

nr=30;

c=500;

r=300;

rv = linspace(1,r,nr);

cv = linspace(1,c,nc);

[C,R] = meshgrid(cv, rv);

before the loop, because they don't change in loop iterations. There is no need to create them in each iteration.

.

[Dinuka Kankanige]

Thank you. This is helpful.

[Star Strider]

As always, my pleasure!

Answer 2

[Matt J]

It is wasteful here to use meshgrid. Also, you should hoist C and R out of the the loop since they don't change with i:

    %resampling
    %new matrix column & row size
    nc=50;nr=30;
    %current matrix column & row size
    c=500;r=300;
    [C,R]=deal(1:(c-1)/(nc-1):c,1:(r-1)/(nr-1):r);
   

for i=1:200
    %loading files into a struct
    A(i)=load(strcat(filepath,filestem,num2str(i),'.mat'));
   
    downscaled_vod{i}=interp2(double(A(i).vod_monthly),C,R);
 
    figure(i)=imagesc(downscaled_vod{i});
end

[Dinuka Kankanige]

Thank you. This too is helpful.