Curvature in NI Vision has a property that makes a lot of sense: "If the current point is too close to either end of the array to choose the additional points, the curvature is calculated as 0." (Vision Manual).
Too close refers - obviously - to ~0.5 Kernel size.
This makes no sense when I'm working on an contour that is "closed" (starting point = ending point) - for example, when I am trying to analyse a particle and its "turning points".
I'm losing 1 kernel width of data at exactly the starting point/end point - as marked in the picture - and in this synthetically generated and exxagerrated case, I'm losing the information about one edge!
To fix this, I either rotate the ROI or change the search direction, calculate the missing data and replace the values in the curvature profile. (Or - calculate the curvature myself.)
This makes absolutely no sense.
Vision could easily recognize starting-point = ending-point, or just allow me set some boolean if there is a reason not to make this automatic. (I can't think of one.)
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Any idea that has received less than 4 kudos within 4 years after posting will be automatically declined.