Color Image Restoration Research

The aim of scientific observation is to acquire the true scene of objects. Limited by many factors in physical principle and/or in capabilities of instrument as well as the effect of noise, however, we can obtain only a degraded image of objects. The processes of image acquisition, transmission, transformation, and application are usually accompanied with a degradation of image. Image recovery encompasses a class of techniques of estimating images, on the basis of degraded observations. The principal content of the techniques includes deconvolution and the estimation of images, and suppression of noise. In actual life, we often encounter the degraded color images, and these images are usually accompanied with color distortion. For example, the biomedical images. This paper’s research work is expand around the color digital pathological slide image’s recovery and correction.In this thesis the differences among the three restoration algorithm were compared, Including inverse filtering, wiener filtering and the improved constrained self-adaptive image restoration algorithm, error-parameter analysis was then described that is fit for accurate blur identification. The reasons of serious ring effect in real blur image restoration are analyzed. Comparing the suppressing ring effect results of recycle edge method and optimized window method, finally the combining wiener filtering and recycle edge method is adopted to restore blur image in the paper. To solve the color distortion problem in color digital pathological slide images, this paper discusses the "Signal Mapping Law" and the "Group Mapping Law" and the " Group Mapping Law" is adopted for color correction. About the algorithm implementation, using VS2010and MATLAB2010a for fast arithmetic speed. Finally, make a graphical user interface to process color correction for convenience. The experimental results show that we can recovery and correct color images effectively with fast arithmetic speed.