I want to process houghlines under any circumstances (at night and at day (winter/summer...))
I'm using simple mean threshold algorithm on grayscale histogram for sobel edge detection.
I tried using histogram equalization before this process (because I thought it would be better), but the edge detection isn't better, but worse.
I'm considering to switch to canny edge detection on color picture, which has been histogram equalized
or should I simply replace my simple mean threshold algorithm with Bernsen's method?
I'm sticking with histogram equalization, because I think, with my camera results, histogram equalization ensures the most realistic pictures (compared with eye view...) (no dark areas, no overexposed areas, etc.)
My pictures exists of right half "dark" target & left half open surroundings.
what is the best approach here?
thanks in advance & best regards.
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
