Research On Color Interpolation And Image Sharpening Algorithms Based On Bayer Pattern Digital Image Sensors

In recent years, with the development of digital imaging technology, digital cameras and other digital products are becoming increasingly popular. This dissertation researches into the color interpolation algorithms and image sharpening for digital cameras.In order to get the full-color images, we need to calculate the other two missing color components by the pixels around the known pixels. Image sharpening is one of the basically ways to enhancee the edge of the image.Considering the character of the distribution of Bayer pattern image, this paper presents a new color interpolation algorithm method based on the bilinear and edge detection. A new color interpolation algorithm method is a step-by-step interpolation method that combines the green component and red & blue components to reconstruct color components. By using the image data characteristics, and the introduction of the triangle edge detection algorithm, the horizontal direction, vertical direction,45 degree and 135 degree directions are considered as image edge detection, not only the image edge detection can be effectively improved, but also the involved of the pixels number are reduced, the processing time is also be reduced.Considering the character of the image edge is not obvious, high-frequency components areweakened after the interpolation algorithms. A new sharpening algorithm method is proposed by which make full considered of the time domain characteristics of image based on edge detection. In a digital camera shooting images, select two successive frames of image, first, we need to take the dynamic testing by the sum of the absolute deviation of SAD, and using improved Sobel operator for edge detection, through the first two of the operation, the dynamic changes in the edge part are detected, then considering the current frame, using improved Laplace operator template, called the spatial filtering sharpening, for the no dynamic change edge portion, considering both the two frame, using the time domain filtering properties and the improved Laplace operator template called time-domain sharpening. The restored images are obvious improved in detail, edge and edge noise. At the same time, compared with the conventional sharpening algorithms, the proposed method has the advantages of low computational complexity.