Correct idea, not quite the correct approach.
Try this:
t= 0:.01:1;
AA = cos(t);
BB= sin(t);
Lv = AA > BB; % Logical Vector
figure
plot(t,AA)
hold on
plot(t,BB)
patch([t(Lv) fliplr(t(Lv))], [AA(Lv) fliplr(BB(Lv))], 'r')
patch([t(~Lv) fliplr(t(~Lv))], [AA(~Lv) fliplr(BB(~Lv))], 'g')
hold off
producing:
.This uses ‘logical indexing’ to separate the two regions.
The independent variable (‘t’ here) may need greater resolution for more complicated curves for this approach to work correctly. If the curves are data, use linspace to create the indpendent variable vector with greater resolution, and interp1 to interpolate the data to it.
