刊期:双月刊
主管单位:中国科学院
主办单位:中国科学院动物研究所,中国昆虫学会
地址:北京市朝阳区北辰西路1号院5号中国科学院动物研究所
邮编:100101
电话:010-64807137
传真:010-64807137
E-Mail:entom@ioz.ac.cn
刊号:ISSN 2095-1353
        CN 11-6020/Q
国内发行代号:2-151
国际发行代号:BM-407
发行范围:国内外公开发布
定价:138元/册
定价:828元/年
银行汇款:中国工商银行北京海淀西区支行
户名:中国科学院动物研究所
帐号:0200 0045 0908 8125 063

您所在位置:首页->过刊浏览->2021年58卷第3期



迁飞昆虫生物学参数反演及种类辨识分析
Recent developments in radar technology that allow the identification of migratory insects
王 锐 张 帆 胡 程 孔少洋 李卫东
点击:803次 下载:29次
DOI:10.7679/j.issn.2095-1353.2021.058
作者单位:北京理工大学雷达技术研究所,北京 100081;北京理工大学前沿技术研究院,济南 250300
中文关键词:昆虫雷达;昆虫RCS;参数反演;机器学习;种类辨识
英文关键词:Radar Research Laboratory, Beijing Institute of Technology, Beijing 100081, China; Advanced Technology, Beijing Institute of Technology, Jinan 250300, China
中文摘要:
虫害严重威胁着我国的粮食安全,迁飞昆虫中有许多是农业害虫,其远距离迁飞是导致虫害异地暴发的重要原因。昆虫雷达是观测昆虫迁飞最有效的工具,在迁飞昆虫的监测和预警中发挥着越来越重要的作用,但传统昆虫雷达不能准确获取昆虫的各项生物学参数,因此无法实现昆虫种类的精确识别。随着雷达技术的创新和发展,通过昆虫雷达获得较为准确的昆虫生物学参数成为可能,为基于昆虫雷达实现迁飞昆虫个体种类辨识提供了依据。本文综述了从雷达回波中提取多频和极化散射参量,然后基于不同的电磁散射反演昆虫生物学参数的方法,并对比分析了基于不同方法的昆虫体重、体长、体宽和振翅频率的反演精度。最后基于生物学参数,采用5种机器学习算法以高精度实现了23种迁飞昆虫的种类辨识,并分析了昆虫生物学参数的测量误差对迁飞昆虫种类辨识精度的影响,初步验证了利用雷达实现高精度迁飞昆虫种类辨识的可行性。
英文摘要:
Insect pests are a serious threat to food security in our country. Many migratory insects are agricultural pests, and their capacity for long-distance migration can cause, devastating, often unexpected, outbreaks of these pests. Insect radar is the most effective tool for observing insect migration and is playing an increasingly important role in monitoring the migration of insect pests and providing early warning of outbreaks. However, because traditional insect radar cannot accurately estimate various biological parameters it cannot accurately identify species. Recent innovations and developments in radar technology, however, make it possible to obtain sufficiently accurate biological parameters to reliably identify migratory insect species. This article reviews the methods of extracting multi-frequency and polarization scattering parameters from radar echoes and summarizes ways of deriving insect biological parameters from different patterns of electromagnetic scatter. It also compares and analyzes the accuracy of determining insect weight, body length, body width and wing-beat frequency based on different methods. Finally, the performance of five machine learning algorithms used to identify 23 migratory insect species, and the influence of measurement errors on the accuracy of species identification, is assessed and discussed. This review demonstrates the feasibility of using radar to achieve high-precision identification of migratory insect species.
读者评论

      读者ID: 密码:   
我要评论:
版权所有©2024应用昆虫学报》编辑部 京ICP备10006425号
本系统由北京菲斯特诺科技有限公司设计开发
您是本站第9020940名访问者