图像识别算法在中草药鉴别中的应用  

Application of Image Recognition Algorithm in Identification of Chinese Herbal Medicine

作　　者：曾辉 沈红岩[2] 张宝文 陶佳[2] 董秀秀 ZENG Hui;SHEN Hong-yan;ZHANG Bao-wen;TAO Jia;DONG Xiu-xiu(Baoding Plant Protection and Quarantine Station,Baoding 071000,China;Agricultural University of Hebei,Baoding 071000,China;Agrculture and Animal Husbandry Bureau of Guangping County,Guangping 057650,China)

机构地区：[1]保定市植保植检站,河北保定071000 [2]河北农业大学,河北保定071000 [3]广平县农牧局,河北广平057650

出　　处：《河北农业科学》2018年第2期100-101,共2页Journal of Hebei Agricultural Sciences

基　　金：保定市哲学社会科学规划课题(2017283)

摘　　要：引入图像识别算法,以颜色特征和纹理特征相结合的方法实现中草药显微图像的识别。针对传统局部二值模式的不足以及K近邻法的多优解问题,加入邻域点灰度值间的关系,对所有的相邻像素进行二值化;以Chi平方统计法对直方图进行相似性度量;结合训练集与测试集,验证优化算法在中草药横截面显微图像识别方面的效果。The image recognition algorithm is introduced to realize the recognition of Chinese herbal medicine microscopic images by combining color features and texture features. In view of the shortcomings of the traditional local binary pattern and the multi optimal solution problem of K nearest neighbor method,the relationship between the gray values of the neighborhood points is added to the two values of all adjacent pixels,the similarity of histogram is measured by Chi square statistics method,and the training set and test set are combined to verify the effects of cross-sectional microscopic image recognition of Chinese herbal medicine.

关 键 词：中草药图像识别 显微图像 局部二值模式 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]