Hi,
I'm using HoughLinesP function to detect lines in an image.
I can manually find the "correct" value to my purpose, to a specific image. but its hard for a set of images, so I wanted to do some run time calculation for the threshold. I have some ideas to achieve this : 1.define the value to zero, but use only the half of the most 'strongs' lines.
2.calculate the average value in the image, and define it as the threshold.
Do they seems logical to you ?
Do anyone know about some solution to it ?
(I've done a search, but want to hear as much of ideas as i can..)