基于可见—近红外光谱技术的家蚕蚕种鉴别方法的研究  

DISCRIMINATION OF VARIETIES OF SILKWORM EGG BASED ON VISIBLE-NEAR INFRARED SPECTRA

作　　者：黄敏[1] 何勇[1] 黄凌霞[2] 楼程富[2] 

机构地区：[1]浙江大学生物系统工程与食品科学学院,浙江杭州310029 [2]浙江大学动物科学学院,浙江杭州310029

出　　处：《红外与毫米波学报》2006年第5期342-344,359,共4页Journal of Infrared and Millimeter Waves

基　　金：国家自然科学基金项目(30270773);高等学校博士学科点专项科研基金资助课题(20040335034);高等学校优秀青年教师教学科研奖励计划(02411);浙江省自然科学基金人才基金资助项目(RC02067)

摘　　要：提出了一种结合主成分分析和人工神经网络技术的可见-红外光谱家蚕蚕种快速鉴别新方法.主成分分析法用于家蚕蚕种品种的聚类分析及主成分的提取.从主成分1和2对所有建模样本的得分图可以看出,主成分分析法对不同种类家蚕蚕种具有较好的聚类作用,可以定性分析家蚕蚕种品种.提取了6个能解释原始光谱的大部分信息的主成分,作为BP神经网络的输入,建立了三层BP人工神经网络模型.选取了4个典型的家蚕蚕种品种,共120个样本,其中随机选取了100样本用来建立神经网络品种鉴别模型,对未知的20个样本进行预测,结果表明,品种识别准确率达到100%.说明该方法具有很好的分类和鉴别作用,为家蚕蚕种的品种鉴别提供了一种新的途径.A new method which was based on principal component analysis （PCA） and artificial neural network （ANN） was developed to discriminate the varieties of silkworm eggs nondestructively by visible and near infrared spectroscopy （ Vis/NIRS）. Principal component analysis （ PCA ） was used to analyze the clustering of silkworm egg samples, and offered the principal components of silkworm egg samples. The score plots of first and second components show that PCA can provide the reasonable clustering of the varieties of silkworm eggs, and can be used to analyze the silkworm eggs varieties qualitatively. The scores of the first 6 principal components computed by PCA were applied as the inputs of a back propagation neural network with one hidden layer. 100 samples from four varieties were selected randomly to build BP-ANN model, and then the model was used to predict the varieties of 20 unknown samples. The discrimination rate of 100%＂ was achieved. It indicates that this model is reliable and practicable. So this model can offer a new.approach to the fast discrimination of varieties of silkworm egg.

关 键 词：近红外光谱 蚕种 主成分分析 人工神经网络 聚类 

分 类 号：S881[农业科学—特种经济动物饲养] TH744.1[农业科学—畜牧兽医]