摘要
不变矩作为理想的形状描述子,具有平移、旋转、缩放变换不变的特点。以M.K.HU矩方法为基础,提出一种采用不变矩的矢量面目标匹配方法。该方法以7个矩不变量组成矢量面目标的特征向量,通过度量矢量面目标与候选面目标特征向量之间的欧式距离判定同名实体。对两种不同数据源中水域面目标进行匹配实验,验证该方法的有效性。
Invariant moment is a perfect shape descriptor,which has the invariant characteristic under transformation(translation,rotation,scale).Based on the method of M.K.HU moments,a matching method of vector area object using invariant moments is proposed in this paper.In this method,the eigenvector of vector area object consists of seven moment invariants,and the same entity is judged through measuring the Euclidean distance between the eigenvector of vector area object and the eigenvector of candidate vector area object.A similarity matching experiment between the water area objects from different data sources is made,and the experiment results validate the proposed method.
出处
《测绘工程》
CSCD
2017年第2期5-8,13,共5页
Engineering of Surveying and Mapping
基金
国家自然科学基金资助项目(41201391
41571399)
关键词
形状描述子
不变矩
匹配
面目标
同名实体
shape descriptor
invariant moment
matching
area object
same entity
