How to smooth 2d matrix

47 次查看(过去 30 天)
varunesh
varunesh 2016-5-29
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

回答(2 个)

Walter Roberson
Walter Roberson 2016-5-29
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.
  1 个评论
varunesh
varunesh 2016-5-29
I don't understand that what are you saying? can you write this in a code. please help me

请先登录,再进行评论。


Kristoffer Walker
Kristoffer Walker 2020-7-21
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
  1 个评论
Christian Ballesteros
编辑:Christian Ballesteros 2021-5-10
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;

请先登录,再进行评论。

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by