Harris Corner Detection Algorithm Based on Improved Contourlet Transform

摘要

According to multi-resolution analysis theory, this paper constructed a new Harris multi-scale corner detection algorithm based on contourlet transform. The algorithm can be decomposed by contourlet under different scales, and extract feature points on the edge direction, obtain the corner. The new algorithm can also overcome phenomenon caused by the single-scale Harris corner detection such as the loss information of corners, the corner position offset and extract false corners disturbed by noise. Experimental results show that the new corner algorithm detected more uniform distribution and reasonable, can be well applied in many fields such as image stitching.