Remote Sensing Data Fusion Image Quality Evaluation Studies

In recent years, with the rapid development of remote sensing technology, remote sensing image fusion become a hotspot in the filed of remote sensing technology. What the multi-source-remotely-sensed image fusion technique study is how to use the image originating form different kinds of sensors synthetically, and achieve a new image with more precise and more abundant information through a certain algorithm. Before the fusion image output must be evaluated. The correct image quality evaluation is a meaningful but very difficult problem. It has a very important position in the image processing. Have a reliable image quality evaluation method, can correctly appraise the image quality, processing technology of the pros and cons and the discretion of the performance of the system. At present, although put forward various evaluation methods, but it is lacking in a unified of theory and method. Now, it is no measure of image quality evaluation. In this paper, it mainly studies the basic methods of pixel level fusion, and sets up a comprehensive image evaluation system. It is used in experiment which is in this paper.Firstly, in this paper, it introduces the meaning and aim of remote sensing data fusion. It introduces the basic method of image processing too. Secondly, it introduces the traditional pixel level fusion method. It is including brovey fusion; intensity hue saturation transform fusion and principal component analysis transform fusion. In this paper, it explained their fusion principle and fusion process. It has a simple evaluation in these three image fusion method. Then, the quantitative evaluation of image is studied. In this part, it is according to the analysis with traditional image evaluation index, setting up a comprehensive image quantitative evaluation system. It is including spectrum fidelity index, image sharpness index, information indexing and structure similarity index. Spectrum fidelity index and image sharpness index is a comprehensive of traditional index. So, there is a total and objective system for image quality evaluation. Information index is sets up with spectrum fidelity index and image sharpness index. Structure similarity integrated human visual characteristics. It is sets up with mean value, variance and covariance. The structure similarity index is sets up with spectrum fidelity index and image sharpness index. Finally, it selects the best combination bands from multi spectral image which come from the satellite-QuickBird. It is used the methods of pixel level fusion in image fusion with panchromatic image and multispectral image. It is used image quality evaluation in the fusion results.Experimental results show that, the image quality evaluation system is feasible which build up in this paper. It can be more comprehensive and objective analysis of fusion image. Combining qualitative and quantitative these two aspects to image quality evaluation. Meanwhile, in this paper, it has analysis the problem with this image quality system.