Method Research On Multi-sensor Image Fusion

Multi-sensor image fusion refers to the techniques that integrate complementaryinformation from multi-image sensor data so that the fusion images are more suitablefor the purpose of human visual perception and computer processing. Each type ofsensor is designed for some given application. There are many of redundancy andreciprocity in the images obtained from the multi-sensor. It is very possible andnecessary to make use of these images to obtain more distinct images and more reliablesystems.In this dissertation we focus on the method of image fusion and evaluation criteriaof its quality, The structure, purpose, application of image fusion are also involved.Firstly, some routine fusion methods are introduced in detail, such as the principalcomponent analysis-based pixel average method and the maximal or minimal pixelselection method. The multi-resolution image fusion techniques which are involved inthis article based mainly on Wavelet transform and Contourlet transform(nonsubsampled Contourlet transform andàtrous-Contourlet transform). Finally, at thelast chapter, the mutli-focus light fusion, remote sensing image fusion and medicalimage fusion are system researched in the multi-scale and multi-resolution transformdomain, and some novel fusion methods (such as New Weighted Gradient SelectionAlgorithm) are developed to determine the coefficients. The results of experiment provethat those transforms and the corresponding proposed algorithms can effectivelyimprove the performance of image fusion system.