Edge-based Image Interpolation Algorithm

Image interpolation is an important basic technology in the digital image processing. Traditional interpolation algorithms are simple and easily realized. They had the same basic principal, which was needed to find an input pixel, weight another pixel nearby, then output the gray-value. The difference is the sequence of getting the pixels around and the number of pixels. Traditional interpolation algorithm has a good visual effect, but they can not process the pixels whose local character of tempestuousness is jump, such as edges, textures, they can lead the edge detail to be fuzzy and jagged in the high-resolution image. For digital video image interpolation, the image is clear also requires not only requires low computational complexity.Based-on the emergency command system, this paper is aimed to solve the problems described above. The main research work in this text is summarized as follows:(1) For the shortcomings of Canny edge detection algorithm, a kind of improved adaptive Canny edge detection algorithm is given which can detect more edge information and improve the ability to denoise.(2)To obtain a clearer image edge, two aspects of multi-directional interpolation algorithm based on the maximum gradient of edge are improved. On the one hand, on the base of simple mean processing of interpolation direction and flat areas of pixels, the impact of its edge pixels is taken into account to fix edge adjacent pixels accurately. On the other hand, the adjacent neighborhood is spreaded to adjacent neighborhood.(3) For video images, adaptive and fast edge interpolation algorithm based on threshold control is improved. All the pixels are divided into flatness and edge class by threshold control, and texture is separated from edge according to the characteristics of the texture image, and then they are interpolated by classification.(4) When multi times magnified the image, the way of cycle-revise by little ratio magnification is selected other than the one-time magnified method to improve the image quality. For RGB combined color images, color fusion was processed first, then use the improved edge optimization interpolation algorithm of this text. By this way, image quality can be improved and computation time can be reduced.(5) Finally, experiments show that the proposed improved algorithms could achieve their desired objectives.