The Research On Fusion Algorithm Of Images Base On Pixel-level

Image fusion technology is an important foundation branch in the field of image processing. It is described by the fusion of multiple sensors in the same single target scene.It used complementary and redundant information to obtain more comprehensive scene description. In order to obtain useful information from the image. Domestic image fusion technology development is slow. There is an urgent need for theoretical research and technology development.This is the pixel-level fusion algorithm based on the image.Fusion image source is a multi-focus image and remote sensing image. Among them, the remote sensing image source is multispectral and panchromatic image. The main contents are as follows:(1) Doing research for pixel-level image fusion. The basic principle of image fusion. It introduces the classical algorithm based on pixel-level image fusion. And the classical algorithm in MATLAB programming. The advantages and disadvantages of each method summary and analysis.(2) Discussion on the quality of the image fusion evaluation. Mainly from subjective and objective analysis. Details of the various parameters and objective evaluation.For comparison later image fusion effect and quality to provide a theoretical basis.(3) In the multi-focus image fusion. In this paper, treated in Space Law.A multi-focus image based on DCT Fusion Algorithm is proposed.It improved integration rules.The algorithm is very simple.To enhance the efficiency of Image fusion.And reduces the computational complexity.Easy to implement in hardware. Finally, after experimental comparison.Comprehensive results figures and tables.Evaluated from the primary objective of both.Obtain better quality image fusion.(4) In the Remote Sensing Image Fusion, using frequency domain method. A novel fusion scheme for multi-spectral (MS) and panchromatic (PAN) image based on compressive sensing and YCbCr space is proposed. The algorithm cited in the compressive sensing signal sampling process is complete compressed signal.Thereby reducing the information transmission and storage of the pressure. This facilitates the processing of remote sensing images as the existence of such a large amount of data problems. In this paper, YCbCr color space conversion method similar to the IHS transform. But based on YCbCr conversion law in regard to IHS transform can better maintain the spectral information. This is very important in remote sensing image. Finally, after experimental comparison. Be considered from subjective and objective. The algorithm is feasible. Image fusion can be enough to maintain spectral information and spatial resolution information.