A Discriminative Image Registration Model with Hybrid Feature Detector

Image registration is a manner of aligning two or more images the same scene which achieves maximum accuracy. Analysis geometric distortion, caused by difference in viewing angle, sensor resolution and distance. The critical step image collection feature points evaluate spatial transformation mainly when outliers are present. In our proposed framework we have used hybrid feature-based area-based methods together. Before applying those methods, Bilateral Filter applied for preprocessing both images, reference that needs to be registered order make noise free. For detection, ORB (Oriented FAST Rotated BRIEF) basically fusion key point detector BRIEF descriptor with many modifications enhance performance. KNN (K-Nearest Neighbors) matching similar reducing miss matches efficient algorithm Random Sample Consensus (RANSAC). template method matching.