Research Of Vegetation Segmentation Technology In Harbin City

Vegetation plays a tremendous and particular role in energy conversion and matter cycling of the earth, the relationship between vegetation and environment is an important aspect of vegetation study. As the resource of the earth surface system, acquisition of dynamic changes information is of great importance to reveal regional and global ecological environment evolving regulars. Nowadays researching vegetation on the earth surface by remote sensing technology and GIS methods are important approaches to research the grass land planning. Vegetation information in remote sensing image is reflected by spectral characteristic and spectral variation of foliage and vegetation canopy.This paper used MATLAB as the platform, and used aerial photograph of Harbin as the research object, according to the principle of color segmentation, collected a lot of sample points of vegetation in different photographs to analyze the distribution regularity and distribution range of the RGB value of vegetation sample points, and decided the discriminant condition of vegetation in aerial photographs. This article used two methods to segment aerial photograph. One is extracting vegetation by discriminant condition directly, the other is using k-means algorithm combine with color discriminant condition to extract vegetation automatically. Both of these two methods could segment the vegetation in aerial photographs effectively, but they are different. This paper have comparative analyzed the two segment methods.