利用局部不变特征识别复杂平面多边形

Recognizing Planar Polygon by Local Invariant Features

提出一种基于最佳匹配点、利用交比来识别平面多边形目标的方法 该方法利用目标的顶点构造交比不变量 ,并以其作为目标的特征矢量 采用最大隶属度原则 ,首先通过局部联合矢量寻找最佳匹配点和判断目标是否丢失顶点 ,然后识别目标 实验表明 :该方法容易实现 ,计算量小 ,识别效率高 。

Optimal matching point is used as the datum point to recognize planar polygons Five adjacent vertices of the polygon are used to construct cross ratio, which makes up the object′s eigenvector Local unit vector is used to seek optimal matching point It also can judge the object′s incompleteness or losing vertex Experimental results show that the method has little computational complexity and stable performance, especially for the local incomplete objects