I have a very simple Mat object which is a binary image. All the pixels are 0 except a few which are 1(some of which are inliers and some are outliers). I want to fit a line on these 1 pixels using RANSAC. Is there any built in method to do that or I have to write one?
I am using OpenCV v2.4.8 on Fedora 20 GNU/Linux.
You probably want to use Hough Lines:
http://docs.opencv.org/doc/tutorials/imgproc/imgtrans/hough_lines/hough_lines.html
FooBar, I have considered HoughLines. Unfortunately it was not giving me what I wanted. Later I attacked my original problem in a different approach which does not require either Hough fitting or RANSAC. But I plan to write a RANSAC line fitting function later in my free time.
According to this SO answer, the OpenCV FitLine function already does this.
Asked: 2014-10-31 14:23:43 -0600
Seen: 12,767 times
Last updated: Dec 06 '14
Hi. Are you using C++, java, python... ?
FYI, here are the different classifiers available in OpenCV http://docs.opencv.org/trunk/modules/ml/doc/ml.html
I know that the function findHomography is able to use RANSAC to find the a good fit between two sets of points, but I personnaly don't know if you can use it as a standalone without ripping it form the source code of findHomography/other relevant source code.
Luckily someone knowledgeable will answer!
Thank you Doombot :) btw, I am using C++