How do I count the number of times my graph crosses two separate lines?

I have this graph:

I want to count the number of times that the graph goes from y=-0.08 (the bottom line) and y=0.16 (the top line). If it goes up and down inbetween those two lines, I don't care. I just want it to count every time it hits -0.08 and goes back to 0.16. For example, in the graph below, count should equal 3 (1 between 18.5 and 19 seconds; 1 between 19.5 and 20 seconds; 1 between 20.5 and 21 seconds; and it shouldn't count the tiny maximum peak at about 19.75 because the graph hadn't yet reached the bottom line again)

(I've seen this done with just a minimum or just a maximum but I don't know how to get it to only count from the bottom line to the top line)

[Dyuman Joshi]

Please share what you have attempted yet.

Answer 1

[Mathieu NOE]

hello

maybe this ?

the second plot shows the valid points (Ncounts = 4)

% dummy data
n = 5000;
x = 50*(0:n-1)/n;
y = -0.3*sin(x).*(1+ 0.75*cos(x/4))/2;

% counts signal crossing from negative (threshold1) to positive (threshold1)
% only consecutive negative to positive crossing points are valid
threshold1 = -0.08; % 
threshold2 = 0.16; % 

[t0_pos1,t0_neg1] = find_zc(x,y,threshold1);
[t0_pos2,t0_neg2] = find_zc(x,y,threshold2);

% plot all crossing points 
figure(1)
plot(x,y,'b',x,threshold1*ones(size(x)),'--r',x,threshold2*ones(size(x)),'--g',...
    t0_pos1,threshold1*ones(size(t0_pos1)),'dr',t0_pos2,threshold2*ones(size(t0_pos2)),'dg','markersize',12);grid on
title('Signal plot');
legend('signal','lower threshold crossing points','upper threshold  crossing points');
xlabel('time (s)')
ylabel('Amplitude')

% keep only valid points 
% n_top = numel(t0_pos2);
n_low = numel(t0_pos1);
Ncounts = 0;

    for ck = 1:n_low
        if ck <= n_low-1 % look for valid upper crossing points (first existing upper crossing point between 2 lower crossing points)
            ind = find(t0_pos2>t0_pos1(ck) & t0_pos2<=t0_pos1(ck+1),1,'first');
        else % maybe we also have a last upper crossing point after the last lower crossing point (must be counted as well)
            ind = find(t0_pos2>t0_pos1(ck),1,'first');
        end
      
      if~isempty(ind)
          Ncounts = Ncounts+1;
          t0_pos1_valid(Ncounts) = t0_pos1(ck);
          t0_pos2_valid(Ncounts) = t0_pos2(ind);
      end
      
    end

figure(2)
plot(x,y,'b',x,threshold1*ones(size(x)),'--r',x,threshold2*ones(size(x)),'--g',...
    t0_pos1_valid,threshold1*ones(size(t0_pos1_valid)),'dr',t0_pos2_valid,threshold2*ones(size(t0_pos2_valid)),'dg','markersize',12);grid on
title('Signal plot');
legend('signal','lower threshold valid crossing points','upper threshold valid crossing points');
xlabel('time (s)')
ylabel('Amplitude')

Ncounts

Ncounts = 4

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [ZxP,ZxN] = find_zc(x,y,threshold)

% put data in rows
x = x(:);
y = y(:);

% positive slope "zero" crossing detection, using linear interpolation
y = y - threshold;
zci = @(data) find(diff(sign(data))>0);  %define function: returns indices of +ZCs
ix=zci(y);                      %find indices of + zero crossings of x
ZeroX = @(x0,y0,x1,y1) x0 - (y0.*(x0 - x1))./(y0 - y1); % Interpolated x value for Zero-Crossing 
ZxP = ZeroX(x(ix),y(ix),x(ix+1),y(ix+1));

% negative slope "zero" crossing detection, using linear interpolation
zci = @(data) find(diff(sign(data))<0);  %define function: returns indices of +ZCs
ix=zci(y);                      %find indices of + zero crossings of x
ZeroX = @(x0,y0,x1,y1) x0 - (y0.*(x0 - x1))./(y0 - y1); % Interpolated x value for Zero-Crossing 
ZxN = ZeroX(x(ix),y(ix),x(ix+1),y(ix+1));

end

[Mathieu NOE]

A second test case where I simply changed the sign of y data :

here we have 3 valid counts

% dummy data
n = 5000;
x = 50*(0:n-1)/n;
y = +0.3*sin(x).*(1+ 0.75*cos(x/4))/2;

% counts signal crossing from negative (threshold1) to positive (threshold1)
% only consecutive negative to positive crossing points are valid
threshold1 = -0.08; % 
threshold2 = 0.16; % 

[t0_pos1,t0_neg1] = find_zc(x,y,threshold1);
[t0_pos2,t0_neg2] = find_zc(x,y,threshold2);

% plot all crossing points 
figure(1)
plot(x,y,'b',x,threshold1*ones(size(x)),'--r',x,threshold2*ones(size(x)),'--g',...
    t0_pos1,threshold1*ones(size(t0_pos1)),'dr',t0_pos2,threshold2*ones(size(t0_pos2)),'dg','markersize',12);grid on
title('Signal plot');
legend('signal','lower threshold crossing points','upper threshold  crossing points');
xlabel('time (s)')
ylabel('Amplitude')

% keep only valid points 
% n_top = numel(t0_pos2);
n_low = numel(t0_pos1);
Ncounts = 0;

    for ck = 1:n_low
        if ck <= n_low-1 % look for valid upper crossing points (first existing upper crossing point between 2 lower crossing points)
            ind = find(t0_pos2>t0_pos1(ck) & t0_pos2<=t0_pos1(ck+1),1,'first');
        else % maybe we also have a last upper crossing point after the last lower crossing point (must be counted as well)
            ind = find(t0_pos2>t0_pos1(ck),1,'first');
        end
      
      if~isempty(ind)
          Ncounts = Ncounts+1;
          t0_pos1_valid(Ncounts) = t0_pos1(ck);
          t0_pos2_valid(Ncounts) = t0_pos2(ind);
      end
      
    end

figure(2)
plot(x,y,'b',x,threshold1*ones(size(x)),'--r',x,threshold2*ones(size(x)),'--g',...
    t0_pos1_valid,threshold1*ones(size(t0_pos1_valid)),'dr',t0_pos2_valid,threshold2*ones(size(t0_pos2_valid)),'dg','markersize',12);grid on
title('Signal plot');
legend('signal','lower threshold valid crossing points','upper threshold valid crossing points');
xlabel('time (s)')
ylabel('Amplitude')

Ncounts

Ncounts = 3

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [ZxP,ZxN] = find_zc(x,y,threshold)

% put data in rows
x = x(:);
y = y(:);

% positive slope "zero" crossing detection, using linear interpolation
y = y - threshold;
zci = @(data) find(diff(sign(data))>0);  %define function: returns indices of +ZCs
ix=zci(y);                      %find indices of + zero crossings of x
ZeroX = @(x0,y0,x1,y1) x0 - (y0.*(x0 - x1))./(y0 - y1); % Interpolated x value for Zero-Crossing 
ZxP = ZeroX(x(ix),y(ix),x(ix+1),y(ix+1));

% negative slope "zero" crossing detection, using linear interpolation
zci = @(data) find(diff(sign(data))<0);  %define function: returns indices of +ZCs
ix=zci(y);                      %find indices of + zero crossings of x
ZeroX = @(x0,y0,x1,y1) x0 - (y0.*(x0 - x1))./(y0 - y1); % Interpolated x value for Zero-Crossing 
ZxN = ZeroX(x(ix),y(ix),x(ix+1),y(ix+1));

end

[Annelise]

For some reason, matlab says that find_zc(x,y,threshold) is an unrecognized function/variable.

I think it's not working because my data isn't a nice, defined line, but a bunch of closely spaced data points from and ecg heart rate

[Annelise]

I'm confused on what find_zc does, so I can't find another way to code it. I can't really find documentation on it. I'm assuming my problem is that my data is discrete points, instead of an equation?

[Annelise]

Nevermind, I didn't see the function definition at the end. This works, thank you so much!

[Mathieu NOE]

as always, my pleasure !

Answer 2

[Jonas]

maybe too caomplicated, but this can work:

data=[1:10 9:-1:1 2:10 9:-1:1 2:10]; % row vector to not get problems with num2str later

lowThr=3;
highThr=6;

% remove data between those thr as those are not interesting

data(data>=lowThr & data<=highThr)=[];

% create 1 if it is under lowThr and 2 of it is bigger than highThr

asBin=(data<lowThr)*1 + (data>highThr)*2

asBin = 1x26

1 1 2 2 2 2 2 2 2 1 1 1 2 2 2 2 2 2 2 1 1 1 2 2 2 2

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

% to string without whitespace

asStr=erase(num2str(asBin),' ')

asStr = '11222222211122222221112222'

% remove duplicates
asStr=regexprep(asStr,'1+','1')

asStr = '122222221222222212222'

asStr=regexprep(asStr,'2+','2')

asStr = '121212'

% if you want to count pairs of "under threshold and then over threshold"
numel(strfind(asStr,'21'))

ans = 2

% or vice versa
numel(strfind(asStr,'12'))

ans = 3

[Annelise]

My data is imported from an excel spreadsheet, with an x and y input. I like the way you solved this, I don't see how I would apply it to the formatting of my data.

[Jonas]

the question is if you also want to know the location at which the crosses appear. if you are only interested in the number of times it crosses, then you dont need your x data, you can just read your y data (e.g. readmatrix) from the excel sheet