基于最小惯性轴及链码的图像形状描述方法  

Shape description based on axis of least inertia and chain

作　　者：李宗民[1] 陆天波[2] 桑鑫焱[1] 秦宝山[3] 

机构地区：[1]中国石油大学(华东)计算机与通信工程学院,山东东营257061 [2]国家计算机网络应急技术处理协调中心,北京100029 [3]北京邮电大学,北京100876

出　　处：《通信学报》2009年第4期1-5,共5页Journal on Communications

基　　金：国家重点基础研究发展计划(“973”计划)基金资助项目(2004CB318000,2007CB311100);国家自然科学基金资助项目(60533090)~~

摘　　要：提出了基于最小惯性轴及链码的结合方法,这种方法能够同时利用形状边界轮廓和区域信息,并利用由特征点和形状质心构成的特征三角形计算得出的三角隶属函数值作为重要特征值进行相似性计算。此方法对于形状的转换是不变的,对凹边形匹配是健壮的,通过实验对比,此方法具有较高的检索性能。The imaged could be seemed to the approximation of the polygons, so it was important to the image match that researched the polygon match method. It is based the combination of the axis of least inertia and the chain, this method was capable of preserving both contour as well as region information, and it utilized the degree of triangular membership which computed by the feature triangular to compute the similarity between the two objects. This method was invariant to image transformations, and robust to concave polygons. The experimental results show that this method＇s performance and retrieval efficiency is well.

关 键 词：最小惯性轴 多边形 顶点 质心 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]