The Research And Application Of Adaptive Mathematical Morphology Based On Hypergraph

Mathematical morphology,based on rigorous mathematical theory,has a simple and efficient basic idea and has been widely used in various fields of image processing.Mathematical morphology uses structural elements to analyze and process the structural features of images.Structural elements of different shapes and sizes can lead to great differences in the results of mathematical morphology operations.In the traditional mathematical morphology,the shape and size of structural elements are fixed,which often do not conform to the local characteristics of some areas of the image,resulting in unsatisfactory image processing results.Therefore,automatic selection of appropriate adaptive structural elements based on image structural features plays a very important role in improving morphological processing results.In order to solve the problem of losing image structural features caused by fixed structural elements,a new adaptive mathematical morphology model based on hypergraph is proposed in this paper,which combines adaptive morphology and hypergraph theory.In this paper,a new gray-scale adaptive mathematical morphology model based on hypergraph is proposed,which combines gray-scale adaptive mathematical morphology with the basic theory of hypergraph.The application of the new model in edge extraction and filtering of gray-scale image is also introduced.Based on the existing gray-scale adaptive morphology theory,the gray-level pixel similarity is defined to measure the similarity of different pixels in gray-level image,and the method of automatic threshold selection is given.The mapping method from gray-level image to hypergraph is defined,and the method of adaptive selection of structural elements is given.A new gray-level adaptive morphology operator based on hypergraph is proposed.At the same time,the detail change rate of gray-level image is introduced to evaluate the degree of detail loss of gray-level images before and after morphological operations.The correctness and validity of the new operator are analyzed and verified by experiments.Experiments show that compared with the existing gray-level morphological operators,the new operator can retain the fine structure features of the image more completely.Based on the new operator,a gray-scale adaptive morphological edge extraction operator and a filtering operator are proposed.Experiments show that the new edge extraction operator can extract the small edges more completely in the image,and the filtering operator has better denoising effect and better preservation of the structural features of the image.Then,according to the characteristics of color image,the new operator is extended to color image.A new color adaptive mathematical morphology model based on hypergraph is proposed,and the application of the new model in edge extraction and filtering of color image is introduced.Combining with the characteristics of color image itself,the similarity of color pixels is defined to measure the similarity of different pixels in color image,and the method of automatic threshold selection is given.The mapping method from color image to hypergraph is defined,and the method of adaptive selection of structural elements is given.A new color adaptive morphological operator based on Hypergraph is proposed.The concept of color image detail change rate is introduced to evaluate the degree of detail loss of color image before and after morphological operation.The experimental results show that the new operator has lower detail change rate and can avoid the loss of image texture details while preserving the color information of the image completely.Based on the new operator,color adaptive morphological edge extraction operator and filtering operator based on hypergraph are proposed.Experiments show that the new operator has better edge extraction effect and filtering effect.