How to smooth 2d matrix

I have 2d matrix. size is give below. I want to smooth this but smooth function is not working properly with imagesc. please help me. Thank you in advance

[ch]      =  imagesc(all_time,height,smooth(data,1000));
all_time  =  30509x1
height    =   533x1
data      =   533x30509

Answer 1

[Walter Roberson]

smooth() is only defined for a vector argument. Internally it reshapes the input to a column vector. If you want to smooth column by column, you will need to do that in a loop before you display the data.

[varunesh]

I don't understand that what are you saying? can you write this in a code. please help me

Answer 2

[Kristoffer Walker]

Use the ":" operator to convert the matrix to a vector, use smooth() to smooth, and return the content to the original matrix format again using ":" operator. For example

A=[0 0 0; 0 0 0; 0 1 0; 0 0 0 ; 0 0 0]
A(:) = smooth(A(:),3)

No need for fancy conv2, filter, or other commands.

Kris

[Christian Ballesteros]

This way you are not accounting for the effect of side-by-side (columns) samples and you smooth last data of a column with the beginning of the following one, which might not be correlated at all. A proper matrix smoothing requires a 2D filtering window.

Example:

% Input data
[P,T] = meshgrid(0:359,0:180);
A     = awgn(sind(T).*cosd(P),5);       % My matrix to smooth
w     = 5;                              % Size of the sliding window (same number of cols and rows in this case)
% Extrapolate values for current window
[Nr,Nc] = size(A);
Nextra  = 0.5*(w-1);
Ap      = interp2(1:Nc,1:Nr,A,-Nextra+1:Nc+Nextra,(-Nextra+1:Nr+Nextra).','makima');    % 2D extrapolation must use 'spline' or 'makima' interpolation
% Smooth data with sliding window
H  = ones(w)./w^2;                      % The 2D averaging filter
B  = filter2(H,Ap,'valid');             % The smooth resulting matrix
% Visualize data
figure; pcolor(A); caxis([-1 1]); shading interp;
figure; pcolor(B); caxis([-1 1]); shading interp;