sum only consecutive 1's in matrix

5 views (last 30 days)
C.G.
C.G. on 20 Oct 2021
Commented: Matt J on 20 Oct 2021
I have a matrix of data where I want to go down each column in the matrix, and sum the occurences where 1's occur consectively. E.g. if I have
1
0
1
1
0
1
1
I want this to be recorded as 1,2,2 as the groups of 1's are separated by 0's. I want this data to be stored in a new matrix where the sum data for each column is stored in a column (in the same format as the orignal data).
Is this possible?

Accepted Answer

Image Analyst
Image Analyst on 20 Oct 2021
This will do it:
% Read in data.
s = load('PNM.mat')
particleNotMoved = s.particleNotMoved
% Prepare a matrix to hold our output -- the run lengths.
[rows, columns] = size(particleNotMoved)
output = zeros(floor(rows/2), columns)
% Loop over all columns, getting the run lengths of each run.
for col = 1 : columns
% Use regionprops in the Image Processing Toolbox.
props = regionprops(logical(particleNotMoved(:, col)), 'Area');
% Get the run lengths for this column.
runLengths = [props.Area]';
% Insert result for this column into our output array.
output(1:numel(runLengths), col) = runLengths;
end
  3 Comments
Image Analyst
Image Analyst on 20 Oct 2021
Matrices need to be rectangular. They can't have ragged bottoms. You have to have something there, like zeros or -1's or NaNs. You could have a cell array and then have nulls/empty in there. The cells would still be there but you'd see [] in the cell instead of 0 or nan.
How are you going to use these? If you don't need to store them, then runLengths has just the numbers and you can use it immediately. Otherwise if you're using it later, after the loop, It should not be a problem to extract the column you're interested in and get just the numbers:
col = 8; % Whatever one you want
thisColumn = output(:, col); % Has trailing zeros.
runLengths = thisColumn(thisColumn > 0) % Has no trailing zeros.

Sign in to comment.

More Answers (1)

Matt J
Matt J on 20 Oct 2021
See,
Tools for Processing Consecutive Repetitions in Vectors
[~,~,runlengths]=groupLims(groupTrue(particleNotMoved),1)
  2 Comments
C.G.
C.G. on 20 Oct 2021
it says that my input must be a vector but im using a matrix and want it to go down the columns
Matt J
Matt J on 20 Oct 2021
An easy modification.
X=particleNotMoved;
X(end+1,:)=0;
[starts,~,runlengths]=groupLims(groupTrue(X(:)),1);
[~,G]=ind2sub(size(X),starts);
result = splitapply(@(x) {x}, runlengths,G )

Sign in to comment.

Products

Community Treasure Hunt

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

Start Hunting!