摘要
植物的数量分类的主要依据是植物的外观特征,通过提取大量特征数据进行聚类分析获得结果。传统做法都是手工测量采集原始数据,效率较低。由于外观特征都可以以数字图片方式获得,通过计算机图像处理分析等技术采集数据并做聚类分析将大大提高效率。关键问题在于特征自动分析和获取,以植物叶片为例,阐述了如何提取大小、叶形及叶缘特征的方法,改进了圆形度参数的定义。提出了计算机辅助植物识别(CAPI)的概念,并对其前景做了讨论和展望。
Plant quantitative classification is conducted on the basis of the visible characteristics of plants.The result is usually come from the clustering analysis of many characteristic data.Traditionally,the data are collected manually and inefficiently.Since the characteristics can be obtained in the forms of numerical images,the efficiency of clustering analysis can be improved with computeraided image analyzing techniques.The key solution is to obtain and analyze the characteristics automatically.The article illustrates the methods of obtaining the size,shape and edge characteristics of leaves and modifies the definition of circular parameter.The concept of comptueraided plant identification is put forward and its prospect is discussed.
出处
《浙江林学院学报》
CAS
CSCD
北大核心
2003年第3期281-284,共4页
Journal of Zhejiang Forestry College
基金
浙江省教育厅资助项目(20020980)
