Information Fusion Based on Optical Flow Field and Feature Extraction for Solving Registration Problems

A novel information fusion method for solving the automated registration problems was proposed. The main information sources include the global optical flow field and feature extracted from images. Our primary contribution is an improved estimation algorithm for computing the optical flow field using anisotropic diffusion. Moreover, we show that the registration process can be consolidated through the background registration and moving target registration that estimates global deformation while ensuring robustness to systematic errors such as those caused by moving foreground objects or occlusion. The validity and accuracy of the algorithm of optical flow on synthetic and real data are demonstrated. The simulation experiments for infrared images show that the fusion method is ideally suited for the application of automated image registration