摘要
根据昆虫图象,对半翅目、鳞翅目、鞘翅目的28种昆虫提取的形状参数、叶状性、球状性等7项数学形态特征进行了主成分分析,利用主成分分析法所得出的结果与传统分类方法相比较,结果表明:半翅目中除黑哎猎蝽独立外,基本上归属于鞘翅目中,鳞翅目中除棉铃虫独立外,其他昆虫种类仍在一类。这说明主成分分析的分类结果与传统分类有较大的相似性,同时也揭示了传统的分类结果可能存在不足,利用数学形态特征和先进的数学方法对昆虫分类进行研究具有重要的意义。
According to insect images, 7 items of math-morphological features such as shape parameter, lobation and sphericity extracted from the images of 28 species belonging to the Hemiptera, Lepidoptera and Coleoptera were conducted for principal components analysis, and the analysis results were compared with traditional classification method. The results showed that the insects in Hemiptera belonged to Coleoptera basically, except that Eectomcorerpa atrgx was independent. In Lepidoptera, except that Helicoverpa armigera was independent, the other insects were in the same category yet. This indicated that the classification results of principal components analysis were more similar to that of traditional classification and also revealed that results of traditional classification might have some defects. Using math-morphological features and advanced mathematical method to study insect classification had important meaning
出处
《安徽农业科学》
CAS
北大核心
2007年第34期11054-11055,11071,共3页
Journal of Anhui Agricultural Sciences
关键词
昆虫分类
计算机视觉技术
目级阶元
主成分分析
Insect classification
Computer visual technique
Class elements on the orders grade
Principal components analysis
作者简介
樊金献（1963-），男，河南桐柏人，讲师，从事生物学教学工作。
