利用翅的数学形态特征对蛾类昆虫进行分类鉴定的系统研究Ⅱ——在科级阶元上的应用  

Systematic research on the classification of moths using math-morphological characters of wingsⅡ:Family level

作　　者：蔡小娜[1,2] 高灵旺[2] 黄大庄[1] 王志刚[1] 沈佐锐[2] 

机构地区：[1]河北农业大学林学院,河北保定071000 [2]中国农业大学农学与生物技术学院,北京100193

出　　处：《中国农业大学学报》2013年第5期96-103,共8页Journal of China Agricultural University

基　　金：河北省自然科学基金项目(C2012204008);国家948项目(2013-4-75)

摘　　要：在科分类阶元上以鳞翅目灯蛾科、刺蛾科、枯叶蛾科、天蛾科、舟蛾科和夜蛾科6科33种蛾类昆虫为研究对象,提取其前后翅图像数学形态特征即偏心率、球状性、叶状性、似圆度、矩形度、延长度和不变矩Hu1、Hu2等共计26项,分析这些特征参数在科级阶元上分类的可行性、可靠性和重要性。结果表明,筛选出6项特征参数作为科级阶元分类鉴别的指标,它们作用的大小依次为(FW矩形度、HWHu2和HW似圆度)>FW似圆度>(FWHu6和FWHu5),回归判别和交叉判别结果的正确率分别为100%和87.9%。Thirty three species of moths belonging to six familys （Arctiidae, Eucleidae, Lasiocampidae, Sphingidae, Notodontidae, Noctuidae） were used in this research of math-morphology at the family level. Twenty six math- morphological characters （MMCs） of both wings were extracted, such as eccentricity, sphericity, lobation, roundness, rectangularity,elongation,and movement invariants including Hul and Hu2. The classificatory importance was analyzed for each MMC. The results indicated that six MMCs could be selected as the classification variables at family level. The contribution of these variables were ranked as the order of （FW-rectangularity, HW-Hu2, HW-roundness） 〉 FW- roundness〉（FW-Hu6, FW-Hu5）. The accuracies of regression and intersecting discriminant analysis therein were 100. 0% and 87.9%, respectively.

关 键 词：蛾 翅 数字鉴定 亲缘关系 科阶元 

分 类 号：Q969.42[生物学—昆虫学] G254.1[文化科学—图书馆学]