Study Of The Color Tree Image Segmentation

With the development of computer image processing technology and continuous improvement, based on the theory of the "precision forestry" a series of related research contents appear in succession, the tree image segmentation for forest precision pesticide spraying n the visualization reconstruction of image、to assess the growth status of the trees and so on lay the basis for the research, this paper study the segmentation of tree image segmentation method for color, software, the main segmentation method is:To make a over green feature transformation、gradient transform、K-mean clustering of transformation, make them as the input image of watershed improvement method, then do the opening and closing reconstruction operation, extract the target and background image for the result image, make the distance transformation only based on the target and the background image (internal and external constraints) exist under the condition of minimum area, at last, do watershed segmentation on it.To make first segment of K-mean clustering, the maximum entropy theory is introduced, respectively using two-dimensional maximum entropy and gray-gradient maximum entropy to clustering’s results for segmentation, combine morphological processing, compare the segmentation results.The original color image is converted to gray image, then build C-V model, do the level set curve evolution movement, to do the morphological processing for the segmentation results, so achieve the purpose of image segmentation, at the same time, by the analysis of the segmentation results in different conditions, find out what are the factors affecting the segmentation result.All kinds of methods are integrated into the GUI experiment system, make the experiment system convenient.Through the experimental compare the various image segmentation methods for results of color image trees segmentation, proved the feasibility of tree image segmentation.