Research On The Optimization Of Digital Image Interpolation Algorithm

A series of image interpolation algorithm are developed in recent few years, which are used to improve the image performance and visual effect. All of these methods focus on only two fields including edge-preserving and interpolation method. Edge information is a main factor which determines the visual effect. It is also the key for target identification and image matching. Thus, the digital edge-preserving is becoming a focus in recent studies.A lot of traditional image interpolation technique is studied in order to obtain better visual effect and reduce the computation. As an algebraic interpolation method, the Newton interpolation is applied in this study. A low-resolution image is enlarged to a series of high-resolution figures by use of this technique. The image performance is improved without updating any equipment. Meanwhile, the cost of bandwidth is saved during the image transmission. It is demonstrated that this interpolation technique decrease the cost of computation and the complexity.In order to eliminate detail fuzzy and the edge jaggy, an algorithm was proposed based on the edge gradient deblurring algorithm. By applying double-channel technique, the detail's blur and alias of images are eliminated. By checking the original image edges, the edge neighbors are amended by a self-relativity interpolation algorithm based on multi-direction gradient variety. The magnifying of image was taken by utilizing the little ratio cycle-revise. Experimental results show that the algorithm can effectively avoid blurring and artifacts edges and generate high quality images with less data, which decreases the computational complexity. If this image is RGB, color fusion is processed first. Then image edges are detected. And the visual quality of the color images more prominent.The new interpolation algorithm is prominent process in several kinds of image such as the natural scenery, the animal image and the character image. A great deal experiment data and image processed have confirmed this article interpolation method feasibility and the validity.