Searching for structures in signal

I have a filtered signal were the base line is zero. I wish to search this signal for structures, i.e., find all the values in between the zeros, and store these induvidual structures as individual vectors for later comparison. These structures are not of equal length. How do I do this efficiently?

Answer 1

[Star Strider]

Try this —

t = linspace(0,25, 500);                                            % Create Data
signal = sum(sin(0.5*rand(5,1)*2*pi*t+rand(5,1)*pi));               % Create Data

zxi = find(diff(sign(signal)));                                     % Indices Of Approximate Zero-Crossings
tlen = numel(t);
for k = 1:numel(zxi)
    idxrng = max(1,zxi(k)-2):min(zxi(k)+2,tlen);                    % Index Range For Interpolation
    ti(k) = interp1(signal(idxrng), t(idxrng), 0);                  % 'Precise' Zero-Crossings
end

for k = 1:numel(ti)-1
    tc{k} = t(t>=ti(k) & t<ti(k+1));                                % Time Vector Between Zero-Crossings
    sigc{k} = interp1(t,signal,tc{k});                              % Associated Signal Vector
end

figure
plot(t, signal)
hold on
plot(t(zxi), signal(zxi), 'sr')
hold off
grid
title('Original Signal & Approximate Zero-Crossings')

sps = numel(tc);
for k = 1:sps
    subplot(ceil(sps/3),3,k)
    plot(tc{k}, sigc{k}, 'LineWidth',1.2)
    grid
    title(sprintf('Segment #%02d',k))
end
sgtitle('Signal Segments Between Zero-Crossings')

.

[Sebastian Daneli]

I will, thx.

[Star Strider]

As always, my pleasure!