›› 2003, Vol. 46 ›› Issue (2): 201-208.

• 研究论文 • 上一篇    下一篇

数学形态学在昆虫分类学上的应用研究.Ⅱ.在总科阶元上的应用研究

赵汗青1, 沈佐锐1*,于新文2   

  • 出版日期:2003-04-20 发布日期:2003-04-20

Use of mathmorphological features in insect taxonomy. Ⅱ. At superfamily level

ZHAO Han-Qing1, SHEN Zuo-Rui1**, YU Xin-Wen2   

  • Online:2003-04-20 Published:2003-04-20

摘要: 从总科角度对鳞翅目(Lepidoptera)和鞘翅目(Coleoptera)5个总科23种昆虫图像中提取的昆虫面积、周长等11项数学形态特征进行了统计分析。结果表明,在总科阶元上,11项特征的可靠性大小为圆形性>(面积、周长、横轴长、球形性、似圆度、偏心率)>(纵轴长、叶状性)>(形状参数、亮斑数)。从数学形态特征角度讲,夜蛾总科等3个总科的亲缘关系远近大小依次为夜蛾总科与凤蝶总科>蚕蛾总科与凤蝶总科>夜蛾总科与蚕蛾总科。

关键词: 数学形态特征, 昆虫分类, 计算机视觉技术, 总科阶元

Abstract: Statistical analyses of 11 math morphological features (MMF) (such as area, perimeter, etc.) from images of 23 species of insects from the Lepidopter a and Coleoptera families Noctuoidea, Bombycoidea, Papilionoidea, Scarabaeoidea and Chrysomeloidea indicates that the ranked reliability of MMF in the identification of insect superfamilies is, from high to low: circularity>(area, perimeter, X-length, sphericity, roundness, eccentricity)>(Ylength, lobation)>(form factor, hole number). From the perspective of mathematical morphology, the kinship of three superfamilies can be ranked as follows: Noctuoidea & Papilionoidea>Bombycoidea & Papilionoidea>Noctuoidea & Bombycoidea.

Key words: Mathmorphological feature(MMF), insect classification, computer vision technology, superfamily