It would help to have the actual data.
x = 1:4;
y1 = rand(size(x));
y2 = rand(size(x));
xi = linspace(min(x), max(x), numel(x)*10); % Increase 'x' Resolution
y1i = interp1(x,y1,xi); % Increase 'y1' Resolution
y2i = interp1(x,y2,xi); % Increase 'y2' Resolution
zxi = find(diff(sign(y1i-y2i))); % Appriximate Indices Of Intersections
for k = 1:numel(zxi)
idxrng = max(1,zxi(k)-1) : min(numel(xi),zxi(k)+1); % Index Range For Interpolation
xix(k,:) = interp1(y1i(idxrng)-y2i(idxrng),xi(idxrng),0); % 'x' Value At Intersection
yix(k,:) = interp1(xi(idxrng),y1i(idxrng),xix(k)); % 'y' Value At Intersection
end
Intersections = table(xix,yix)
figure
plot(x, y1, x, y2)
hold on
plot(xix, yix, 'sm')
hold off
grid
This will be reliable if there is only one intersection between each pair of indices. It may fail to discriminate intersections that are close together, even with the increased point resolution.
EDIT — (6 Apr 2022 at 13:00)
Made ‘xi’ length a function of the original ‘x’ length.
.
