The Reaserch On Digital Image Processing Technology Based On Interpolation

Image interpolation is an important task in the field of digital image processing. It is fundamental to many digital image operations, such as scaling, rotation, and geometric correction.The discrete and continuous viewpoints of image interpolation, the general steps of image interpolation,and space transformation in image interpolation are studied first.Shannon information theory is described in detail. Through the analysis of image sample model and sample distortion, we sum up the confusion sampling aperture distortion and distortion that brings to the image quality, and comparatively study of the simulation models commonly used in image interpolation The classical linear interpolation algorithms are investigated systematically. Through the analysis of traditional linear interpolation algorithm, we analyze traits of zoomed images produced by these linear interpolation techniques, and find their common defect. The adaptive interpolation methods developed in past decades are also studied. Their advantages and disadvantages are discussed.Adaptive interpolation techniques produce more vivid zoomed images compared with traditional interpolation techniques, but they are difficult to be used in actual image processing system for their huge computational burden. On the basis of thorough study of traditional WaDi interpolation algorithm, WaDi image interpolation based on local gradient was proposed, which is a space-variant technique. The algorithm considers the overall characteristics of the image, fully fexcavates the space gradient information and local asymmetric features so as to increase high frequency components due to sampling. The edge components of the image have been strengthened and can obtain clear image details. This algorithm can be applied to image scaling and rotation and the magnification can be integer, rational and irrational, thus solving some of the problems in practical application. It is simple and easy to realize.In view of the sensitivity of human eye to edge of image, error amending theorem is introduced into image interpolation in the paper. An adaptive edge-preserving algorithm for image interpolation based on error amending is proposed. We discuss the error amending theorem, edge detection and classification method at length. A large number of simulation experiments in this paper validate that the two methods can achieve preferable results to some image data such as the standard image, nature, photographs, and other types of text images.