Jeru’s answer is correct for this case. If you have a case with bigger noise where morpholigocal operations won’t take them out, you can make a cutoff with the contour size, something like for contour in contours if cv2.contourArea(contour) > minimal_length before counting 1 solved How to calculate the number of shapes detected after thresholding
Best way to visualize a dependent variable as a function of two other independent variables, each of them is a column of a datafarme? solved Best way to visualize a dependent variable as a function of two other independent variables, each of them is a column of a datafarme?
A possible approach. Find the convex hull of these points. Then find the two vertices which are the furthest apart. This gives a first diagonal. Then find the two other vertices that are the farthest from the first diagonal. This gives a second diagonal. Now you can just use the quadrilateral formed by these four … Read more