ABSTRACT
The application areas of multi-sensor image fusion (MSIF) are battlefield monitoring, remote sensing, machine vision, robotic vision, airfield/airport surveillance, enhanced vision system and medical imaging. The MSIF can be performed at three different levels: pixel level, feature level and/or decision level. This chapter presents pixel-level-based MSIF using multi-resolution discrete cosine transform (MDCT) along with results. A simple pixel-level MSIF is to take the mean of the grey-level registered source images pixel-by-pixel. The chapter aims to develop a novel MDCT-based image fusion algorithm to fuse the registered source images. In the image registration process, the information in the source images is adequately aligned and registered prior to merging of the images. In image processing discipline, discrete cosine transform (DCT) is an important and widely used transform for researchers and technologists. In this transform domain, large DCT coefficients are concentrated in the low-frequency region and it is known to have good energy compaction properties.
