摘要
目的以人工神经网络和模式识别技术为基础,对动物饲料中常添加的违禁西药进行了自动识别与分类。方法对违禁药物的显微图像进行预处理、图像分割,并提取纹理特征。根据提取的特征应用BP人工神经网络完成对5种违禁药品(呋喃妥因、呋喃唑酮、呋喃它酮、氯霉素、扑热息痛)的识别分类。结果该算法不仅能快速识别出上述5种违禁药品,且准确度比较高。结论该方法对于以上5种违禁药品能得到满意的识别结果。
Objective To identify and classify illegal medicine in fodder automatically based on artificial neural network and pattern recognition techniques. Methods Preprocessing, image segmentation, feature extraction of microscopic images of medicine, and BP artificial neural network were used to complete the recognition and classification of furantoin, furazolidone, furahadone, chloramphenicol, paracetamol. Results The results show that this method can quickly identify the illegal medicine with high accuracy. Conclusion For the 5 kinds of medicine,this method could achieve a satisfactory recognition results.
出处
《广东药学院学报》
CAS
2013年第6期631-635,共5页
Academic Journal of Guangdong College of Pharmacy
基金
广东省科技计划项目(2011B040300031)
关键词
BP神经网络
模式识别
特征提取
纹理特征
违禁药物
显微图像
BP artificial neural network
pattern recognition
feature extraction
texture feature
illegal drugs
microscopic images
