Research On Image Edge Enhancement Based On Additive Wavelet And Morphological Gradient

In the technology of digital image processing, edge detection present an important content of the image analysis, theoretically, with the image edge information, all the information can be recoverable. Therefore, edge detection can be seen as the key to dealing with many problems, it is widely used in target recognition, image segmentation, remote sensing, medical image analysis and other fields, while plays an important role in image processing.Mathematical morphology is a mathematical method to analyze the geometry and structure, on the basis of algebra, which is to be seen the science of quantitative description by the math logic. Its main idea shows that use of the geometric template called the structural elements to detect the objects’shape in the image. Mathematical morphology is as a nonlinear analysis method, has become a major research area in image processing, and its theory and method have indicated a powerful tool in the field of image analysis. This paper described the basic theory of mathematical morphology and some of its basic operations, such as corrosion, expansive, open, closed, etc., and also studied the research of their basic properties, characteristics and its applications in image enhancement.The image edge enhancement presents the important content of image processing and analysis. This article described some method of image enhancement commonly used, overviewed the definition of the image edge enhancement, method, application, meaning, as well as domestic and foreign research status, analyzed several operators about classical edge detection, for example:Roberts operator, Prewitt operator, Sobel operator, Laplacian operator, the Log operator and Canny operator. The paper also analyzed their different and application in image edge extraction comparatively.By analyzing the deficiencies of the traditional image edge detection techniques, in-depth analysis of the characteristics of the additive wavelet transform and mathematical morphology on the basis, a new image edge extraction method has been proposed by combining with additive wavelet transform and mathematical morphology. Firstly, this method was made on the multi-scale additive wavelet decomposition of the original image; secondly, the low-frequency image was filtered to obtain morphological gradient map by morphological gradient, by using modulus maxima method to extract the edge on each high-frequency sub-image, then overlaid multi layer edge maps, in order to obtain high-frequency edge map; thirdly, inverse transform of additive wavelet have been made on and morphological gradient map and the high-frequency edge map; lastly, the final edge map was got by binarizing the final results. The results of experiment show that the method of edge extraction has many advantages, such as the integrity, multi-directional, shift invariance and speed, etc, it can extract the image edge which gray value changes little, so that ensured the edge information more abundant.