地标点法在夜蛾昆虫数字化分类中的应用研究  

Application of landmarks on digital classification of noctuid moths (Lepidoptera: Noctuidae)

作　　者：蔡小娜[1,2] 张若楠 李阳[3] 李晶 黄大庄[3] CAI Xiao-Na;ZHANG Ruo-Nan;LI Yang;LI Jing;HUANG Da-Zhuang(Department of Basic Courses, Agricultural University of Hebei, Baoding 071000, Hebei Province, China;Department of Basic Courses, Hebei Finance University, Baoding 071051, Hebei Province, China;College of Landscape and Travel, Agricultural University of Hebei, Baoding 071000, Hebei Province, China)

机构地区：[1]河北农业大学基础课部,河北保定071000 [2]河北金融学院基础课部,河北保定071051 [3]河北农业大学园林与旅游学院,河北保定071000

出　　处：《环境昆虫学报》2019年第1期106-112,共7页Journal of Environmental Entomology

基　　金：河北省高等学校科学技术研究项目(QN2017077);河北省林业科学技术研究项目(1507473)

摘　　要：为探讨地标点法在夜蛾昆虫分类中应用的可行性,本文以13种夜蛾昆虫为对象进行数字化分类研究。首先,利用TpsDig 2软件在夜蛾前、后翅上提取12个地标点,再利用多元统计分析对消除非形状因素信息的地标点坐标进行方差分析和判别分析。研究结果表明对夜蛾前翅特征进行判别分析的准确率为88.7%,对夜蛾前后翅特征进行判别分析的准确率为97.9%,利用前后翅特征对13种夜蛾进行数字化分类的准确率明显高于仅用前翅的结果。研究表明利用蛾翅的地标点特征可以实现夜蛾昆虫的数字化分类。To attest the feasibility of landmarks on digital classification of noctuid moths, thirteen noctuid species were taken as example to be classified. Firstly, twelve landmarks form front and hind wings were selected by using software TpsDig 2. Secondly, multivariate statistical analysis was used to remove non-shape variation form landmark coordinates which were applied to variance analysis and discriminant analysis. The data showed that accuracies of discriminant analysis used to characteristics from front wing only and front and hind wings both were 88.7% and 97.9% respectively. So, the latter was apparently higher than the former. The results indicated that landmarks could be used for digital classification of noctuid moths.

关 键 词：夜蛾 地标点 数字化 分类 

分 类 号：Q964[生物学—昆虫学] S433.4[农业科学—农业昆虫与害虫防治]