基于多尺度形状分析的叶形识别系统

叶形自动识别系统使用一种改进了的边界跟踪算法检测叶片边界,应用多尺度形状分析技术有效地提取出规范锯齿长度和曲率尺度空间图像极大值点集两个曲率特征。综合叶片的离心率、似圆率和这两个边界曲率特征来检索叶片数据库,进行形状匹配,实现叶片自动分类。实验结果表明,该系统使用的技术能大幅提高叶形识别的精确率和检索率。

Leaf shape recognition system with multi-scale shape analysis

The leaf shape recognition system could detect a leaf's boundary through using an improved boundary tracing algorithm, and efficiently extract curvature features of normalized teeth length and curvature scale space image maxima set from the boundary by a technique of multi-scale shape analysis. The system retrieved and classified the leaf by shape matching with the samples in a leaves database by eccentricity, circularity and the curvature features. Experimental results indicate that the method can improve the precision and the recall of leaf shape recognition greatly.