A Kind Of Image Scaling Algorithm Study Based On Mathematicai Morphology

With the rapid development of computer technology, the application of the digital image information is more and more widely, digital image processing technology is then got an unprecedented development and more extensive application. Image scaling as a kind of image processing technology, has an important role in practical applications, it can help people get better access to more useful information, mainly used for image display, transmission, image analysis, animation, film synthesis and so on, in the field of production of living, medical systems, military aspects of aviation system it has a wide range of applications. In the subject of "the Study of Tongue Image Acquisition and Analysis of Chinese Medicine" acquisition tongue will need to use different size of images, the collected images need detail parts of amplification in order to better observe the characteristics, which are needed image scaling technology, and which is the starting point of this article.Mathematical morphology is a kind of science established on rigorous mathematical theory, is a kind of new method used in image processing and pattern recognition domain, the advantage of fastly and simply operation make it has excellent performance in the image edge detection, especially the image with noise interference. This article first introduce the theoretical knowledge of image scaling including principles and some classic traditional scaling algorithm, then introduce the basic operations and several typical operational examples in the binary image and grayscale images with mathematical morphology. Considering that human eyes is sensitive to the edge of image and the advantage of mathematical morphology on the edge detection, This paper proposes a kind of image scaling algorithm based on mathematical morphology, the image after processing has better edge details maintaining, the main contents are follows:On the basis of traditional morphological filtering algorithm, taking into account of the impact of the structural elements to morphological operations it proposes a multi-structure complex mathematical morphology filter to do smoothing, it can smooth more noise and the filtering is better.Using multi-structure to improve the traditional morphological edge detection operator can get multi-structural morphological edge detection operator to edge detection, it can detect the clearly edge of image.Considering that the refinement process with single structure has direction, it uses two sets eight directions of structures to refine process to get more accurate edge. Use special bicubic interpolation method by the scaling to the pixels detected locating in the edge, while in the non-edge domain the pixels are using bilinear interpolation method, then compile them to get the target image.Finally, using matlab to do experiment and simulate this algorithm with processing lena figure and tongue image, comparing and analyzing the results with using the nearest neighbor interpolation method, bilinear interpolation method and bicubic interpolation method, verified that this method has an effectiveness of edge preservation.