Design And Implementation Of A Hibernal Tree Automatic Classification System Based On Bayes

The thesis analyzes features of tree system, studies potential aspects which could influence classification results, puts forward factors that should be taken into consideration in hibernal trees classification modeling and then induces these factors into six main aspects of features. The paper, from a high abstract viewpoint, extracts the most related contents and information for classification, sets up hibernal trees classification model, and finally implements the classification model with Bayes, accomplishing a hibernal tree automatic classification system based on Bayes. The thesis puts forward and implements a tree automatic classification system based on Bayes, which object is hibernal trees. The thesis includes:(1) thoroughly analyzing features of hibernal tree from a classification oriented viewpoint, designs HPF, a hibernal plant description file, for hibernal tree, and successfully applies it to subject of research work(2) analyzing flaws of current color representation method, arguing that it can not be well applied for color measurement and calculation. Then the thesis designs a new color description method, proposes and implements color distance and similarity method, and applies the method to the research(3) discussing integrity of trees, and draws out a conclusion that features of trees can not be treated as discrete and independent ones when classifying trees on different level. The paper accomplishes hibernal tree automatic classification system based on Bayes for species, genus, and family levels(4) designing and implementing self-learning for parameters of features in hibernal tree automatic classification system, in order to get best classification effects by self-learning and necessarily adjusting parameters according to actual data(5) accomplishing test on project data, and giving a detailed analysis to the experiment. Experiment result shows the method proposed in the thesis can well solve hibernal tree automatic classification problems.