基于生物光子学的小麦虫害识别模型

WHEAT PEST RECOGNITION MODEL BASED ON BIOPHOTONICS

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.