摘要
由于有机磷农药的挥发性与稳定性易受环境的影响,降低了电子鼻在分类鉴别中的精确度与可比性。在定位传感器响应起始点的基础上,截取有效数据,以保证分析结果的可靠性;通过对平均微分法、积分值2种特征提取方法分析,提出了一种面积斜率比值法的特征提取方法;运用主成分分析(PCA)和BP神经网络对3种方法2种农药进行分类计算,结果表明:面积斜率比值法的分类效果最好。
Due to the effect of significant environment on the volatile and the stability of organophosphorus pesticides ,the accuracy and the comparability of identification are weakened using electronic nose. In order to ensure the reliability of the results of the analysis, effective analyzed data are obtained by finding out the starting point of sensor response. A new feature extraction method named as＂ slope area ratio＂ method is proposed on the basis of the investigation of the integral method and the mean-differential coefficient value method, which are two kinds of popular feature extraction method. Principal component analysis （PCA） and BP neural networks are used to classify the two types of pesticides. The results show that the slope-area ratio method is the most effective method compared with other two feature extraction methods in the classification of organophosphorus pesticides.
出处
《传感器与微系统》
CSCD
北大核心
2009年第9期25-27,共3页
Transducer and Microsystem Technologies
基金
河南省基础与前沿技术研究计划资助项目(092300410039)
