基于区域增长与边缘检测的立木图像信息提取  

Extraction of Information of Standing Tree Images Based on Region Growth and Edge Detection

作　　者：吴晓兰[1] 高林[1] 阚江明[1] 李文彬[1] 

机构地区：[1]北京林业大学工学院,北京100083

出　　处：《科技导报》2009年第22期44-47,共4页Science & Technology Review

基　　金：农业科技成果转化资金项目(2007GB23600468)

摘　　要：为建立自动立木测量系统,实现林木资源调查管理的信息化、自动化和精确化,本研究提出一种基于区域增长和边缘检测的立木图像信息提取方法。通过颜色分量对标定参照物进行提取,以主干基部区域为种子区域对立木主体进行增长,遇到边缘时停止;再利用数学形态学对其进行处理,得到需要的立木主体部分。根据标定参照物所得到的参数,由摄影测量学求得树干的胸径、侧枝粗细、枝下高等信息。研究表明,该方法在提取标定物和树干轮廓方面更为有效,对立木信息的自动提取更加方便精确。The extraction of information of standing tree images provides a fast and precise way tot management of the forest resource inventory, also provides an important basis for pre-harvest assessment. Methods are proposed in this paper to acquire images of standing trees in real-time, to extract relevant features and to achieve fast recognition without destruction. Thus the rich information can be easily preserved and accessed based on intuition. So it provides a basis for the automation of a forest resource inventory management system, and at the same time, for the improvement of the traditional methods in the furest resource inventory and technology, which also reduces the labor intensity in field investigations. A tree measurement system is proposed to achieve the automatic and precise management of forest resources survey, as an enhancement of the traditional survey methods and techniques. A method of extracting the calibration stick by color components is proposed. First, using a tripod to keep the camera in the horizontal plane, i.e. to make sure that the imaging plane is parallel to the middle curve of the stem volume and normal to the ground. The images of the standing tree can thus be taken. Then a region growth algorithm is applied to search the tree contour. Region growth approach is adopted to achieve the image segmentation, where neighboring pixels arc examined and different choices of seeds may have different segmentation results. According to the shape characteristics of the tree trunk, a seed pixel is chosen from the stem base. Then the main trunk is extracted by mathematical morphology. Finally the information of the diameter at the breast height, the diameter of the main branch and the branch height are determined according to photogrammetry and the parameters of the calibration stick. Experimental results show that these methods can effectively segment the calibration stick and the tree contour. It is convenient to obtain accurately and automatically the standing tree intbrmation.

关 键 词：图像处理 区域增长 数学形态学 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]