TABLE OF CONTENTS An image is never an exact representation of the object under observation; it is always corrupted by degradations during acquisition and within the imaging system itself. These include noise, blurring and distortion. Image restoration removes or reduces these degradations. The point spread function (PSF) or the modulation transfer function (MTF) provides a complete, quantitative description of an imaging system and directly characterizes the image degradation within the system and can be used to restore the fine detail in images. The problem is more complicated if the image is also degraded by significant amounts of noise. Restoration techniques attempt to model the degradation and apply the inverse process to recover the original image. They are most effective when the point spread function or modulation transfer function is known and the nature of the blurring and noise are well understood. Geometric distortions can be reversed using inverse bilinear equations and gray-level interpolation.
After reading this chapter you will be able to:
identify the main sources of image noise and discuss their characteristics;
choose appropriate general strategies for minimizing the effects of noise;
discuss the advantages of adaptive filtering;
model image degradation comprising blur and additive noise;
employ suitable values to Wiener filter a noisy, blurred image;
compare the performance of inverse filtering with Wiener filtering;
explain how distortion can be removed from images.
Images can be degraded by a number of different mechanisms, including noise, blurring and distortion. Noise is present because any imaging device must...
