Fast computation algorithm of image registration using increment sign correlation and M-estimators

摘要

Image registration is the alignment of two or more images so they superimposed. Registration is an important problem and a fundamental task in image processing technique. Increment Sign Correlation is based on average estimation of incremental tendency of brightness in neighbouring pixels. A novel and robust statistics as a similarity measure for robust image registration is proposed. The statistics is named as M-estimator because it is based on the Probabilistic models that describe the likelihood of outcomes and used to predict future events. Probability distribution Allocates likelihood to outcomes and often represented by parameterized functions. A basic algorithm that scans, searches and registered over a large scene is represented with a technique of M-estimator is suggested. The suggested algorithm is robust to light variations, shadows, occlusions, noise, scaling and rotation. Many experimental results with real images are simulated and discussed.