ABSTRACT
This chapter is an introduction to methods for perfecting visual and/or metrological image quality. We begin, in the section “Image Perfecting as a Processing Task,” with the formulation of the image perfecting task. Then, in the sections “Possible Approaches to Restoration of Image Distorted by Blur and Contaminated by Noise” and “MMSE-Optimal Linear Filters for Image Restoration,” we present possible approaches to correction of image blur and cleaning images from additive noise, and introduce MMSE-optimal linear filters that perform this processing and are implemented, for the sake of simplification of their design and minimization of their computational complexity, in a domain of certain orthogonal transform, which features fast transform algorithm. In the section “Sliding Window Transform Domain Adaptive Image Restoration,” we extend MMSE-optimal linear filters for working in sliding window, which makes them local adaptive. In the section “Multicomponent Image Restoration and Data Fusion,” we further extend these filters to multicomponent image restoration and consider the issue of fusing image data from different sources. In the section “Filtering Impulse Noise,” we discuss nonlinear filtering methods for cleaning impulse noise in images. In the section “Correcting Image Grayscale Nonlinear Distortions,” methods for correcting image grayscale distortions are addressed, and finally, in the section “Nonlinear Filters for Image Perfecting,” we provide a survey and classification of nonlinear filters for image denoising and enhancement.
