|
|||||
|
|
| 人工神经网络在蚊虫自动鉴定中的应用(英文) | |
| 引用本文： | 李振宇,周祖基,沈佐锐,姚青.人工神经网络在蚊虫自动鉴定中的应用(英文)[J].四川农业大学学报,2005,23(4):411-416. |
| 作者姓名： | 李振宇 周祖基 沈佐锐 姚青 |
| 作者单位： | 1. 四川农业大学,农学院,四川,雅安,625014;中国农业大学,农学与生物技术学院,北京,100094 2. 四川农业大学,农学院,四川,雅安,625014 3. 中国农业大学,农学与生物技术学院,北京,100094 |
| 基金项目： | The National Priority Program of Innovative Technology （863） No. 2002AA243031 and 2003AA2090.Acknowledgments We wish to extent sincere appreciation to Institute of Mi crobiology and Epidemiology, Academy of Military Medical Sciences and QUBIT SYSTEMS Co. for their respectively supporting and providing the photosensor and WfRec system. |
| 摘 要： | 应用光电传感器和瞬时波形记录系统记录了5种蚊虫的翅振波形。结果显示,每种蚊虫的翅振波形为相似的正弦波。蚊虫的翅振频率虽然彼此间存在交叉,但差异明显。因此,通过建立人工神经网络对蚊虫的种类进行分类识别是可行的。研究中分别以蚊虫翅振频率和翅振波形建立人工神经网络,结果发现以翅振频率为特征值的神经网络的识别准确率高。该网络识别的平均准确率为72.67%,最高为89%。 |
| 关 键 词： | 蚊子 翅振频率 人工神经网络 自动鉴定 |
| 文章编号： | 1000-2650（2005）04-0411-06 |
| 收稿时间： | 2005-05-12 |
| 修稿时间： | 2005年5月12日 |
Automated Identification of Mosquito (Diptera: Culicidae) Wingbeat Frequencies by Artificial Neural Network |
|
| LI Zhen-yu,ZHOU Zu-ji,SHEN Zuo-rui,YAO Qing.Automated Identification of Mosquito (Diptera: Culicidae) Wingbeat Frequencies by Artificial Neural Network[J].Journal of Sichuan Agricultural University,2005,23(4):411-416. | |
| Authors: | LI Zhen-yu ZHOU Zu-ji SHEN Zuo-rui YAO Qing |
| Institution: | 1. College of Agriculture, Sichuan Agricultural University, Yaan 625014, Sichuan, China; 2. China Agricultural; University, Beijing 100094, China |
| Abstract: | The wingbeat waveforms of five species of mosquitoes were recorded by a photosensor and a WfRer system. Wingbeat waveforms of the result show that the mosquitoes is analogical sine wave. Although their wingbeat frequencies are overlapped with each other, the diversities of their mean wingbeat frequencies are obvious. Thus, it is possible to construct artificial neural network for classifying the wingbeat frequencies of five species of mosquitoes. Artificial neural network classifiers are respectively built by wingbeat waveform time series and wingbeat frequencies. The most accurate classifier tested is an artificial neural network by using variable of wingbeat frequency. The accuracy is average 72.67% and highest 89%. |
| Keywords: | mosquito wingbeat frequency artificial neural network automated identification |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |