Corner Detection Using SUSAN Operator for Image Registration
- Login to Download
- 1 Credits
Resource Overview
This MATLAB program utilizes the SUSAN operator to detect corner points in images, computes 128-dimensional feature vectors at these corner locations, matches feature vectors between reference and real-time images, and achieves precise image registration to obtain affine transformation parameters.
Detailed Documentation
This MATLAB program implements corner detection using the SUSAN (Smallest Univalue Segment Assimilating Nucleus) operator to identify feature points in images. The algorithm calculates 128-dimensional feature vectors at detected corner points by analyzing local image patterns around each corner. The program then performs feature matching between reference and real-time images by comparing these feature vectors using similarity measures. This matching process enables accurate image registration, which determines the affine transformation parameters (including rotation, scaling, translation, and shearing) required to align the images properly. The implementation includes functions for SUSAN-based corner detection, feature vector extraction using local gradient information, and RANSAC-based feature matching for robust transformation estimation.
- Login to Download
- 1 Credits