基于生物光子学的小麦虫害识别模型  

WHEAT PEST RECOGNITION MODEL BASED ON BIOPHOTONICS

作　　者：史卫亚[1,2] 焦珂珂 王艳娜[1] 宋红霞[1] 梁义涛[1,2] 

机构地区：[1]河南工业大学信息科学与工程学院,河南郑州450001 [2]河南工业大学粮食信息处理与控制教育部重点实验室,河南郑州450001

出　　处：《河南工业大学学报（自然科学版）》2013年第6期100-104,共5页Journal of Henan University of Technology：Natural Science Edition

基　　金：国家863计划(2012AA101608);国家自然科学基金项目(31171775)

摘　　要：使用生物超弱发光技术,分别对正常小麦和含虫小麦的自发光特性进行测量,并使用小麦自发光特性的位置特征、散布特征和形态特征等9个参数构成小麦分类的特征向量,利用BP神经网络设计分类模型,对特征向量进行训练和测试,试验结果表明模型可以正确区分含虫小麦和正常小麦,正确率达到95%,该模型为小麦隐蔽性虫害的检测提供了一种新的思路.We measured the self-illuminating characteristics of normal wheat and pest-damaged wheat by using ultra weak luminescence technology,constructed a wheat classification feature vector from nine parameters of the wheat self-illuminating characteristics, such as position characteristic, distribution characteristic and morphological characteristic,and designed a classification model by using BP neural network to train and test the feature vector. The results showed that the model could discriminate normal wheat and pest-damaged wheat,and the accuracy was 95%. The model provided a new thought for detecting hiding wheat pests.

关 键 词：小麦虫害 超弱发光 神经网络 模式识别 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]