Research On Fusion Image Quality Assessment Of Visual And Infrared Images

Image quality assessment is a very significant part of image processing. It can compare different algorithms or different parameters in the algorithms with image performance assessment, and further choose right algorithm or the parameters for research and analysis. Fusion image quality evaluation is an important part of quality evaluation. Fusion image quality assessment differs from generic distortion image measurement; it not only preserves the fusion of input images, but also should not have false information impacting on next processing. It takes one of the input images as a reference, but should be better than any of them.On the basis of visible/infrared Images, this dissertation has made systematic studies on various aspects as preprocessing of images fusion and assessment in depth. At the same time, based on importance measures and fuzzy integrals, a new assessment method for image fusion quality with human vision is presented. The main achievements of this dissertation are as follows:(1) It has reviewed the proposed fusion methods and assessment algorithms, described the different assessment methods in detail, summarized common image fusion solutions.(2) Proposed a new fusion assessment algorithm base on fuzzy measures and fuzzy integrals. This algorithm regards the fuzzy integral processing as evaluation processing, obtain the final assessment result. Experiment has demonstrated that the new algorithm is available to assess the fusion image quality with anti-noise property, and can combine subjective and objective assessment. It is appropriate for visible/infrared Images fusion assessment.(3) According to the practical research application, a set of software"Visible/ Infrared Image Fusion Assessment Software System"was developed to evaluate the works involved in the paper. This system is consisted of preprocessing module, fusion module, and evaluation module.