Research On Multisensor Image Data Fusion Algorithms

Image fusion is a important aspect of Image processing technology. As one of the important aspect of image fusion, multisensor remote sensing image fusion has attracted many attentions in remote sensing field.What the multisensor remote sensing image fusion t technology research images data of the same region what originating from different kinds of sensors, and achieve a fused image with more precise, more abundant information through a certain algorithm. As a result of this processing, the fusion image is more useful for further image processing tasks such as object detection and recognition, along with many important military and civil applications.This paper firstly introduced the theories of mutlsensor data fusion and remote sensing image fusion, the development situation of remote sensing image fusion in the national and international. Then, this paper deeply analyses classical methods used remote sensing image fusion such as IHS transform and wavelet transform. Studying and analysesing defect of IHS transform, for improving IHS transform aim at this algorithm defect, this paper bring up what based on IHS transform of histogram matches, this algorithm valid reduce spectrum loss of fusion image, and analyses the effects of the wavelet function and the decomposed level on remote fused image deeply, the research provides basis for choosing wavelet function and the decomposed level.This paper research fusion regulations based on pixel and based on region in wavelet transform algorithm, through analyses experimentation data, discovered algorithm based on wavelet region relation coefficient more superiority than anyone. Then , this paper bring up a new image fusion algorithm what combining to based on histogram matches of IHS transform and wavelet region relation coefficient. This algorithm valid combine to IHS transform what can more retain space resolution of source images and wavelet transform what can more retain spectrum information of source images, so the fusion image both more enhance image's spectrum information and maximum retain image's space resolution. Through analyses experiment data, verified this algorithm's validity.