Tree Segmentation from an Image

Tree is a very common object in nature, thus its segmentation provides quite important information for scene 3D reconstruction. In this paper, we present an algorithm for segmenting trees from a complex scene. The trunk and leaf regions of a tree can be individually identified and the trunk structure of the tree can also be extracted. The proposed algorithm is mainly composed of a preliminary image segmentation, a trunk structure extraction, and a leaf regions identification process. We model the extraction of trunk structure as an optimization problem, where an energy function is formulated according to the color, position, and orientation of the segmented regions. We propose an algorithm to minimize this energy function and thus extract the trunk structure of the tree. After obtaining the trunk structure, the leaf regions can then be easily identified by finding those leaf regions located above the trunk regions. This algorithm has been tested on some real images and the results indicated that our algorithm performed well for these images.