基于颜色特征的昆虫自动鉴定方法
Automatic insect identification based on color characters
刘 芳1**；沈佐锐1***；张建伟1；杨红珍1，2
点击：1680次 下载：0次
DOI：
作者单位：1中国农业大学昆虫学系 北京 100094； 2北京自然博物馆 北京 100050
中文关键词：自动鉴定， 蝴蝶， 形态， 颜色， 计算机视觉
英文关键词：automatic identification, butterfly, morphology, color histgram, computer vision
中文摘要：对昆虫实现自动鉴定是将相关专家从大量重复性的鉴定工作中解脱出来的最有效的方法之一。本章文着重论述昆虫自动鉴定系统鉴定的基本原理，开发步骤和使用昆虫自动鉴定系统进行自动鉴定的步骤，基于颜色特征的昆虫自动鉴定系统开发中在图像获取、颜色特征提取方面的难点和相应的解决方案。最后，经过测试开发出的基于颜色特征的蝴蝶自动鉴定系统在对43种蝴蝶的自动鉴定上取得了95.2%准确率。
英文摘要：Automated insect identification provides an effective solution to reduce the burden of routine identification. This paper reviews the basic principle and developing procedures of the automated insect identification system, the processes of the automatic identification, and the difficulties of capturing images and designing color characters we facing in the processes of developing the automated insect identification system based on color characters and their solutions. Based on above mentioned studies, we developed an automated butterfly identification system based on color characters which got 95.2% veracity in 43 butterfly species automatic identification.