danielshaving / CVFeatures_fast_Harris_Corner_Extraction

Harris Conner key points and descriptors extraction by pure numpy

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Harris Corner Kps and Descriptors Extraction

This Harris Corner Kps extraction codes based on the calculation of Heissen matrix of the Gaussian Kernel windows, estimating the corner points by det(M)/trace(M)

R function (Harris response)

  • Images after gaussian filters Pic
  • Second derivatives in x->x direction, x->y direction, y->y directions
  • M is the 2*2 matrix computed from image derivatives: Pic
  • R response = det(M)/trace(M)

Candidate corner points and threshold

  • get coordinates of candidates and their values (the R response point value > threshold)
  • sort candidates
  • set a min_distance and choose the candidates with the minimum distances

Pic

About

Harris Conner key points and descriptors extraction by pure numpy


Languages

Language:Python 100.0%