I'm trying to create the 'detectHarrisFeatures' function of MATLAB in OpenCV. 'detectHarrisFeatures' in MATLAB returns a 'cornerPoints' object. One of the attributes of this object is 'metric'. Can someone explain what this attribute is and how I can calculate it using OpenCV? Or is there any function which can compute similar values?
That would be the strength of the corner response. So if it's 0, it's no corner at all.
This is similar to the cornerHarris function, where the metric would be the value of the various maximums in the dst Mat. Matlab simply does the local maximum suppression and detection for you.
I've normalized the matrix dst and computed local maximums using a threshold. However, the values of the dst matrix at those points are very high compared to the matlab metric. How can I scale down the dst matrix to gives values similar to matlab? Forgive me if this question is very basic, I am new to Image Processing and OpenCV.
Honestly, I have no idea what MATLAB is doing in the particulars. Try taking the log or the square root.
Asked: 2016-05-17 00:49:47 -0600
Seen: 977 times
Last updated: May 17 '16
