Automatic Image Registration Techniques

This book deal with image registration, an essential pre-processing step for computer vision, remote sensing, and medical analysis applications. The automatic image registration process is divided into two main categories: area-based and feature-based methods. Three feature detectors (SURF, HARRIS and FAST) are applied for the feature detection of high-resolution satellite images. This book includes the findings on the accuracy of automatic image registration depending on the number of selected GCPs, the identified features on the images and the spatial distribution of the chosen GCPs over the image area. The ground topography also affects the accuracy of the image registration process. The observations and investigations given in this book have shown that the SURF feature detector performs satisfactorily to achieve the optimum accuracy level for high-resolution IRS images. The projected outcomes of this work are the accuracy assessment of automatic image registration for high-resolution images using different parameters, and this is the novelty of the study in this book.