Several Remote Sensing Image Fusion Algorithms Base On Contourlet Transfrom

With remote sensing technology developing, specially the development of sensor technology, we have more and more available remote sensing image about an certain region, which present all kinds of characteristics of region, such as high spatial resolution, hyper-spectrum, multi-temporal, multi-angle. Multi-spectral image have more spectral information with little spatial details information, but panchromatic image have more spatial details information and little spectral information, so how to integrate a image owning hyper-spectral information and high spatial details information becomes a studying hotspot. The paper employing in the study of fusion about panchromatic image and multi-spectral image has two following works:First, we propose a remote sensing image fusion algorithm based on the local feature of NSCT coefficients. In low-pass band, we selectively add coefficients of low-pass band of panchromatic image on the basis of that of multi-spectral image. In the high-pass band, we select coefficient according to local standard deviation. The experimental results show that the fusion image not only keeps the spectrum information of source image well, but image resolution has been greatly improved compared to other fusion methods.Second, we divide target area into edge area and non-edge area according to spatial distribution characteristics of ground objects. Obviously, edge area is of full spatial information, should be added more information of multi-spectral image, while non-edge area is filled with spectral information, added more information of panchromatic image. Pulse coupled neural networks (PCNN) is a visual cortex-inspired neural network and characterized by the global coupling and pulse synchronization of neurons. By taking the advantage of PCNN, We present a method for fusing the panchromatic and multi-spectral images by combining NSCT and PCNN. Experimental results demonstrate that the proposed method is feasible both keeping spectral information at non-edge area and improving spatial information at edge area.Remote sensing image data has the feature of massive data*but owning complementary and redundancy among data. So image fusion becomes a efficient method of how to reduce redundancy, improve reliability, and the algorithm of fusion play a important role.