摘要
为了提高辣椒种子品种鉴别效率,实现辣椒种子品种快速无损鉴别,研究了一种基于荧光光谱的辣椒种子品种快速无损识别方法。选择同一系列相似辣椒种子的杂交品种卓椒3号、卓椒4号和卓椒5号为研究对象,分别采集辣椒种子的荧光光谱56份,对原始光谱进行Savitzky-Golay(SG)卷积平滑和一阶导数(first derivative,FD)预处理;采用主成分分析(Principal Component Analysis,PCA)方法提取主成分,实现数据降维;采用Fisher判别分析方法建立辣椒种子品种识别模型。结果显示:采用荧光光谱建立判别模型时,卓椒3号辣椒种子的品种识别正确率达到92.9%,卓椒4号、卓椒5号辣椒种子的品种识别正确率均达到100.0%,整体识别正确率达到97.6%;采用荧光光谱的一阶导数光谱建立判别模型时,卓椒3号、卓椒4号、卓椒5号辣椒种子的品种识别正确率均达到100.0%,整体识别正确率达到100.0%。研究结果表明,种子叶绿素荧光光谱结合化学计量学方法能够有效识别辣椒种子品种。
In order to improve the efficiency of pepper seed cultivar identification,realize rapid and nondestructive identification of pepper seed cultivar,a kind of rapid and nondestructive identification method of seeds for pepper cultivar based on fluorescence spectra was studied.Hybrid pepper seeds ’Zhuojiao No.3’ ’Zhuojiao No.4’ and ’Zhuojiao No.5’ were taken as research samples,and 56 fluorescence spectras of each cultivar seed samples were collected respectively.The original spectral data was pretreated with Savitzky-Golay (SG) convolution smoothing method and first derivative (FD).Principal component analysis (PCA) method was adopted to extract principal components and reduce the dimension of the spectral data,the discriminant model of pepper seed cultivars was established by Fisher discriminant analysis.The results showed that when using the original fluorescence spectra to establish discrimination model,the identification accuracy of ’Zhuojiao No.3’ seeds was 92.9%,the identification accuracies of ’Zhuojiao No.4’ seeds and ’Zhuojiao No.5’ seeds were all 100.0%,and the overall identification accuracy was 97.6%.When the discrimination model was established by using the FD spectra,the identification accuracies of ’Zhuojiao No.3’ ’Zhuojiao No.4’ and ’Zhuojiao No.5’ were all up to 100.0%,and the overall identification accuracy was up to 100.0%.The results showed that it was feasible and nondestructive to detect pepper seed cultivars by fluorescence spectra.
作者
李翠玲
姜凯
马伟
王秀
LI Cuiling;JIANG Kai;MA Wei;WANG Xiu(Beijing Research Center of Intelligent Equipment for Agriculture,Beijing 100097,China;National Research Center of Intelligent Equipment for Agriculture,Beijing 100097,China;Beijing Research Center for Information Technology in Agriculture,Beijing 100097,China)
出处
《蔬菜》
2020年第9期9-13,共5页
Vegetables
基金
2019年现代农业产业技术体系北京市创新团队。
关键词
辣椒
种子
品种
荧光光谱
无损检测
pepper
seed
cultivar
fluorescence spectra
nondestructive detection