AUTOMATIC IMAGE REGISTRATION AND MOSAICKING SYSTEM

The main goal of this work is to bring developed registration methods into one automatic image registration system and make them work operationally.

The developed registration system is a full-featured application intended for operational use by beginners as well as by advanced users. Registration may be achieved by one simple click or may be controlled by several parameters. The system contains toolboxes that increase the registration strength using user knowledge.

Three different algorithms for control point extraction are implemented in the system and other methods can be easily added. One of the algorithms uses optical flow ideas to extract the features in both images. The second method uses spectral information of the images and their local wavelet transform modulus maxima to extract a set of control points. The last one uses centers of gravity of the closed contours and other strong edges as control points.

People: Dmitry V. Fedorov, Leila M. G. Fonseca, Charles Kenney, B.S. Manjunath.

This project is supported in part by CAPES, SELPER, Office of Naval Research, China Lake Naval Air Warfare Center and CalTrans.