hello
maybe this ? NB, this is a no Image Processing Tbx solution (as I don't have it) but I wanted to give it a try
for the small video (8 frames if I'm correct) that you posted here are my results (in degrees ) :
angle_left = 81.9921
82.0263
82.4190
81.5316
82.5122
82.5768
81.7864
81.7500
angle_right = -80.5377
-80.8377
-80.6901
-81.5111
-80.6901
-80.9807
-81.3844
-80.8377
the code basically do a gradient of the image , smooth it a bit (I have used this Fex submission : smooth2a - File Exchange - MATLAB Central) , then I do top (red dots) and bottom (yellow dots) boundary of the gradient data (as well as the "mean" curve => green dots)
personnaly I opted to use the top boundary points and taking the first (left) 5 samples (same on the right side) you can use polyfit to estimate the tangent slope - from where you can get the angles , of course. The tangent vector is displayed as the magenta segment
zoom on the LHS of the plot :
zoom on the RHS of the plot :
NB that the results may vary according to the threshold used to define the boundary points
hope it helps !
code :
% Create a VideoReader object for the AVI file
vidObj = VideoReader('4dropletshorter.avi');
figure(1)
axis equal
% Read video frames until available
k = 0;
while hasFrame(vidObj)
k = k +1;
vidFrame = readFrame(vidObj);
frame = double(rgb2gray(vidFrame))/255;
[angle_left(k,1),angle_right(k,1)] = do_the_stuff(frame); % main function
pause(1);
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [angle_left,angle_right] = do_the_stuff(frame)
A = flipud(frame); % to have image displayed with correct y direction
B = abs(gradient(A)); % abs gradient
% smooth the gradient
nx = 10;
ny = 1;
B = smooth2a(B,nx,ny); % Fex : https://fr.mathworks.com/matlabcentral/fileexchange/23287-smooth2a?s_tid=srchtitle
B = B -min(B,[],'all'); % min is now 0
B = B./max(B,[],"all"); % normalize 0 to 1
[y,x] = find(B>0.25); % adjust the threshold !!
[xtop,ytop,xbottom,ybottom,xm,ym] = top_bottom_boundary(x,y);
% figure(2)
imagesc(B);
colorbar('vert');
set(gca,'YDir','normal');
colormap('gray');
hold on
plot(xtop, ytop, '.r'); % TOP boundary
plot(xm,ym,'.g'); % mean curve
plot(xbottom,ybottom, '.y'); % BOTTOM boundary
% tangent vector left side
samples = 5;
p = polyfit(xtop(1:samples), ytop(1:samples),1);
f = polyval(p,xtop(1:samples));
plot(xtop(1:samples),f,'m-','LineWidth',2)
angle_left = atan(p(1))*180/pi;
% tangent vector right side
p = polyfit(xtop(end-samples+1:end), ytop(end-samples+1:end),1);
f = polyval(p,xtop(end-samples+1:end));
plot(xtop(end-samples+1:end),f,'m-','LineWidth',2)
angle_right = atan(p(1))*180/pi;
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [x3,y3,x4,y4,x5,y5] = top_bottom_boundary(x,y)
% based on FEX : https://fr.mathworks.com/matlabcentral/answers/299796-tight-boundary-around-a-set-of-points
%split data into classes and find max/min for each class
class_label = unique(x);
upper_boundary = zeros(size(class_label));
lower_boundary = zeros(size(class_label));
for idx = 1:numel(class_label)
class = y(x == class_label(idx));
upper_boundary(idx) = max(class);
lower_boundary(idx) = min(class);
% mean curve computation :
indy = diff(class)<2; % indice of contiguous y data
[begin,ends] = find_start_end_group(indy);
duration = ends - begin; % in samples
[md,ii] = max(duration);
% keep only mean values with y dist below threshold
if md>2 % let's find contiguous block of pixel that are at least 3 samples long
mean_curve(idx) = mean(class(begin(ii):ends(ii)));
else
mean_curve(idx) = NaN;
end
end
% left_boundary = y(x == class_label(1));
% right_boundary = y(x == class_label(end));
%
% % left_boundary
% x1 = class_label(1)*ones(size(left_boundary));
% y1 = left_boundary;
%
% % right_boundary
% x2 = class_label(end)*ones(size(right_boundary));
% y2 = right_boundary;
% top boundary
x3 = class_label;
y3 = upper_boundary;
% bottom boundary
x4 = class_label;
y4 = lower_boundary;
% mean y curve
x5= class_label;
y5 = mean_curve;
end
