An investigation on feature model of tree crown in natural scene

For the natural scene recognition and moving object segmentation, the segmentation of tree crown could improve the recognition rate and segmentation accuracy of moving objects when they are in strong wind and no buildings. Up to now, many researches mainly focus on the background modeling and texture synthesizing. In this paper, feature modeling of tree crown is discussed and modified by using the technique of remote sensing physics and texture synthesizing, and the feature model of tree crown is given based on the texture of tree crown. At the same time, the method and procedure for modeling and image processing are presented, and it is proved that the feature model is feasible according to the experimental results.