基于PCA-KNN的金线莲种类识别

Species identification of anoectochilus roxburghii based on PCA-KNN

为了能对金线莲品系进行方便准确地识别,提出基于PCA-KNN的金线莲叶片识别方法。通过图像预处理,获得特征较为明显的叶片区域,再提取纹理和颜色特征,进行特征融合,然后采用PCA降低特征维度,提高识别精度,最后通过训练KNN分类器完成分类。以3个品系的金线莲为例进行鉴别试验,结果表明,提出的识别方法与其它方法相比,正确识别率更高,达到98.4%,能准确识别不同种类的金线莲。

In order to facilitate the accurate identification of the strain of anoectochilus roxburghii,a PCA-KNN based identification method was proposed.Through image preprocessing,the blade regions with more obvious features were obtained,then the texture and color features were extracted for feature fusion,and then PCA was used to reduce the feature dimension and improve the recognition accuracy.Finally,the classification was completed by training the KNN classifier.Identification tests of 3 strains of anoectochilus roxburghii were carried out,and results show that compared with other methods,the proposed method can effectively improve the recognition rate up to 98.4%,and it can accurately identify different categories of anoectochilus roxburghii.