|
Bugs in Hough line detection - examples of CV_HOUGH_PROBABILISTIC failures: msg#00688lib.opencv
It appears there is a major bug in the probabalistic hough line detection algorithm/code. I'm using the hough example app from the documentation (opencvref_cv.htm). Running: lines = cvHoughLines2( dst, storage, CV_HOUGH_PROBABILISTIC, 1, CV_PI/180, 80, 30, 10 ); on the example image from the documentation: http://www.vizzion.com/images/test2.jpg Gives: http://www.vizzion.com/images/test2result.jpg i.e. nothing like the result shown in the documentation - many lines missing, and lots of incorrect lines radiating from the top left corner of the image. The problem is even more apparent when running: lines = cvHoughLines2( dst, storage, CV_HOUGH_PROBABILISTIC, 1, CV_PI/180, 10, 10, 10 ) On this 'ideal' test image: http://www.vizzion.com/images/test1.bmp Which gives this result: http://www.vizzion.com/images/test1result.jpg i.e. some lines not detected, some lines incomplete, one line at a slightly wrong angle. Does anyone get the same poor results on these test images? Chris Change settings: http://www.yahoogroups.com/mygroups, select Get Emails (get all posts) Daily Digest (one summary email per day) Read on the web (read posts on the web only)Or Unsubscribe by mailing OpenCV-unsubscribe@xxxxxxxxxxxxxxx |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Something better than Canny: 00688, Carlos Jiménez Saiz |
|---|---|
| Next by Date: | Re: Re: V4L2 Support: 00688, Matthias Jauernig |
| Previous by Thread: | Something better than Cannyi: 00688, Carlos Jiménez Saiz |
| Next by Thread: | Problem when using IPP and starting/stopping threads...: 00688, Philip Gruebele |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |