Study Of Image Segmentation Algorithm Application To Korean Pine Cones And Male And Female Flowers

Korean Pine is our valuable tree, its economic benefits very considerable. In order to better protect the Korean Pine and improve yield of Korean Pine, we should use modern information means for Korean Pine solid quantity's forecast technique. This paper mainly studies about the pine cones and male and female flowers'image segmentation problem. Because in Korean Pine solid quantity's forecast technique, Korean Pine's cones and the number of Korean Pine's male and female flowers and morphological characteristics are the main parameters, these features will directly influence the accuracy of the prediction results last. According to Korean Pine's cones and Korean Pine's male and female flowers'images have different characteristics, the Snake model segment Korean Pine's cones'images and GVF Snake model segment Korean Pine's male and female flowers images. Ultimate we can get about Korean Pine's cones and Korean Pine's male and female flowers'object segmentation image, and provide accurate parameters for Korean Pine solid quantity's forecast techniqueThis paper mainly research content as follows:First, main image segmentation technology has been analyzed deeply in the paper, according to the feature of Korean pine cones and Korean Pine's male and female flowers images, we select suitable image segmentation technology. By the analysis of experimental results, we found activity in active contour model segment object image of Korean pine cones and male and female flowers is good, but still had the shortcomings, so we need to improve Parameter Active Contour model. After analysis, found that influence is the main reason of the segmentation results of the initial contour selection problem and energy function equation of structural problems.Then focus on solving the orientation of initial contour paper has selected the Ant Colony Algorithm to get initial contour and optimize initial contour by consulting amount of literatures and reference. Using the robustness and the global of Ant Colony Algorithm's excellence to improve the Snake model for image segmentation, the obtainable initial contour not only accurate and initial contour is complete nodes, but also reasonable distribute initial contour's nodes. So it can provide precise initial contour for parameters active contour model to segment image.Finally ,using the improved Snake model to do the simulation experiment that is Korean Pine cones'image segmentation, and paper compare results of the experiment validity; Using the improved GVF Snake model again to the Korean Pine male and female flowers image's edge extraction, the simulation experiment results are achieved the project requirements. About specific means of Korean Pine cones and Korean Pine male and female flowers image segmentation are as follows:(1) In the progress of Korean Pine cones image segmentation, according to its background relatively complicated and edge hollow characteristics, the paper proposed the improved Snake model based on Ant Colony Algorithm. It is overcome the shortcomings of traditional Snake model, and finally gain complete target. Simulation experiments show the effectiveness and accuracy of the proposed algorithm.(2) In the progress of Korean Pine male and female flowers image segmentation, because of its characteristic of complicated background and target edge hollow degree larger, paper proposed improved GVF Snake model based on Ant Colony Algorithm. It is overcome traditional GVF Snake model external field shortcomings, finally get complete target. The Ant Colony Algorithm's optimization is used to determine the initial contour closer to target, so that can reduce the iteration number to increase calculation speed, and still can enlarge capture range, the part of hollow can accurate extraction. Simulation experiments show the effectiveness and accuracy of the proposed algorithm.