Getting KeyPoints & Descriptors with harris in Java OpenCV 3.4.x

asked 2019-06-11 07:50:54 -0500

Pika gravatar image

Hey !

I am currently trying to match images through the differents algorithm proposed by OpenCV. I did match thanks to AKAZE algorithm, which is giving pretty good results. However, I believe that for the kind of images I've to deal with, Harris Corner Detection is more appropriate.

Here's my problem : I've been using it a little with OpenCV 2.4.13 or something, with the FeatureDetector.create(FeatureDetector.HARRIS), this was providing me the KeyPoints. But soon after, my group decided to update the lib to 3.4.2.

My question is : How can I detect features and compute the descriptors in order to match images using Harris as detector ?

Note : I did try ImgProc.cornerHarris(), according to the example : https://docs.opencv.org/3.4/d4/d7d/tu... , it displays correctly the corners detected, but does not return me anything, especially the descriptor of these keypoints, that could be used for any matching.

Thanks for considering my issue, and for any help !

edit retag flag offensive close merge delete

Comments

harris corner detection does not produce any descriptors, and maybe even it's the wrong keypoint detector.

what are you trying to achieve with the whole ?

berak gravatar imageberak ( 2019-06-13 02:08:49 -0500 )edit