摘要
番茄苗产生的挥发物易受到病害、虫害、损伤等多种因素影响。该文利用电子鼻系统测试机械损伤番茄苗挥发性物质的变化,通过主成分分析、线性判别分析对4种不同处理机械损伤的番茄苗进行分析,结果表明主成分分析各处理样本间均有重叠,区分效果不理想,线性判别分析各处理样本基本可以分开;用逐步判别分析和BP神经网络对各处理样本进行判别,测试集的准确率分别达到84.4%和93.8%以上,神经网络模型的预测结果更好。该研究可为番茄苗机械损伤快速在线监测提供参考。
The value of E-nose response signals differed with different levels mechanical （0 pricks, 30 pricks, 60 pricks and 90 pricks） damaged tomato plants, indicating that the emission of volatiles by tomato plants changes in response to different degrees of damage. The tomato plants with different levels mechanical damages were classified through principal component analysis （PCA） and linear discrimination analysis （LDA）. The result showed that the electronic nose could distinguish different damaged tomato plant by LDA. However, samples by PCA were overlapped. Stepwise discriminant analysis （SDA） and back-propagation neural network （BPNN） were applied to evaluate the data. The average correction ratios of testing set of SDA and BPNN were 84.4% and 93.8% respectively. The results indicate that it is possible to classify different degrees of damaged tomato plants using e-nose signals.
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2012年第15期102-106,共5页
Transactions of the Chinese Society of Agricultural Engineering
基金
国家自然科学基金项目(30071246)
浙江省自然科学基金项目(Z5100155)
关键词
植物
损伤检测
神经网络
电子鼻
番茄苗
plants, damaged detection, neural networks, electronic nose, tomato
