昆虫图像自动鉴别技术
Advances in the automated insect image identification
陈小琳1,2**；侯新文2；刘成林2；刘晓秋1；张知彬1***
点击：1867次 下载：0次
DOI：
作者单位：1. 中国科学院动物研究所 北京 100101； 2. 中国科学院自动化研究所, 模式识别国家重点实验室 北京 100080
中文关键词：昆虫， 图像， 模式识别， 自动鉴定系统
英文关键词：insect image, pattern recognition, automated identification system
中文摘要：
昆虫是地球上物种多样性最为丰富的生物类群，其物种鉴定任务复杂而艰巨，可靠的
物种鉴定是开展昆虫学工作的重要基础之一。当前，国内外的人工昆虫物种鉴定能力均不能
满足实际需求，因而人们开始不断探索利用计算机自动鉴定昆虫的原理和方法。目前，模式
识别技术的迅猛发展已为昆虫图像的自动鉴定提供可能。文章概述昆虫图像自动鉴定技
术研究的历史与现状，总结主要原理和方法，介绍工作流程，并展望发展前景。
英文摘要：Insect has the richest species diversity on the earth. The task of ins
ect species identification is one of the most fundamental works for many entomol
ogical fields and usually being difficult and complex. At present, reliable inse
ct identification is mainly carried out by taxonomists, but this approach can't
meet practical needs at all. Therefore, many scientists have tried to resolve th
is problem by using modern computing technology. Pattern recognition technology
has developed rapidly and made automated insect image identification feasible. T
he history, underlying theory, general process and the prospects for developing
an automated insect image identification system are briefly reviewed and discuss
ed.