Segmentation Based On Remote Sensing Images Of Forest Vegetation Texture Blue Noise Theory

In recent years, with the rapid development of high-resolution remote sensing technology,high-resolution remote sensing images can be a very good response entities and their spatiallocation information ground. Vegetation is a general term refers to all plant communitiescover the surface, closely related to soil, climate, topography and water conditions and othernatural environmental factors. Vegetation is one of the best reaction ecological regions, is animportant part of the ecosystem, which in the ecosystem is a very important role. Vegetationget information quickly and accurately, and has great significance for urban construction andgreen space planning. However, the complex structure of vegetation and variety, with amulti-scale clustering and spatial distribution of crushing characteristics, so a gooddescription of the classification of forest vegetation characteristics specific scale pixel can getmore reasonable classification results.To achieve this goal, select an appropriate scale to identify vegetation remote sensingimage information is the foundation of the entire algorithm. In the beginning, a measure basedon the theory of blue noise characteristics of the forest vegetation remote sensing imagetexture is used for remote sensing image-scale extraction of forest vegetation, compared withthe traditional method reduces the time complexity of the algorithm. Extract the optimal scaleis very important in the whole study because the optimal segmentation determines the sizeand proportion of the accuracy of information extraction of forest vegetation. To be able tosplit the polygon original forest vegetation remote sensing image display clear boundaries,can not be too broken or too vague.With further research, texture model extraction method based on a combination ofvarious types of texture features are constantly proposed, but most of the models due to thecombination of high complexity method, time cost is too large, and had reduced expressionin the texture on the exact degree requirements. Lack of concise and exact scale structuralelement to achieve the expression of the model forest vegetation segmentation is a majorproblem.This paper presents a theory based on the support of the Blue Noise establish a rapiddetection of forest vegetation texture scale approach. After the detection of regional scaletexture image selection and continuous detection of narrow blue noise characteristics, todetect the use of blue noise characteristics and the scale to determine the scale of structural elements, structural elements of the use of the resulting morphological filtering to achievesegmentation of remote sensing images of forest vegetation zone. According to the specificstructural elements of forest vegetation analysis was then performed on the remote sensingimages do not hit or strike the conversion process, the use of structural element is designed todetect a scale of forest vegetation segmentation extraction.Further, on the basis of these shrubs can be detected according to the scale, set a specificstructural element for shrubs split, so that the more complex remote sensing images canclearly extraction of forest vegetation and shrubs.