基于电子鼻的番茄种子发芽率检测

程绍明; 王　俊; 马杨珲; 王永维; 韦真博

基于电子鼻的番茄种子发芽率检测

Detection of germination rate of tomato seeds by electronic nose

摘要

摘要: 防止种子掺假、以次充好，为快速无损检测高发芽率的种子，该文将不发芽的浙杂809番茄种子以不同比例掺入到发芽率为92.6%的番茄种子中，得到种子的发芽率分别为90%，80%，70%，60%，50%和0等6种比例，并利用电子鼻对其进行分析。结果表明：利用电子鼻可以很好的区分出番茄种子发芽率为90%、80%、50%～70%、和不发芽的4种情况；当种子发芽率为70%、60%、50%时，其图形信息部分重叠，利用电子鼻较难区分开。在主成分分析和线性判别分析的基础上，利用BP神经网络和支持向量机对上述情况进行分类识别，结果表明：两种识别模式的训练集的正确率分别为93.6%和97.4%，预测集的正确率分别为65.2%和72.7%，相对于BP神经网络模式识别，支持向量机预测系统的误差较小，具有很好的预测性能。

Abstract: In order to find out a fast nondestructive examination method for germination rate of tomato seeds, different samples of tomato seeds with six kinds of germination rates were analyzed by electronic nose, and which were classified through principal component analysis (PCA) and linear discrimination analysis (LDA). The result shows that the electronic nose could distinguish the tomato seeds with germination rate of 90%, 80%, 50%－70%and un-germination seeds. However, samples with germination rate of 50%－70% were overlapped. Based on PCA and LDA, BP neural network (BPNN) and support vector machine (SVM) were introduced in the classification. The results showed that the recognition rates for germination rate of tomato seeds by the two methods reached to 93.6% and 97.4% respectively with training set, and 65.2% and 72.7% respectively with forecast set. Compared to BPNN, SVM method has less predicting errors, which has better forecasting performance.

HTML全文

参考文献(0)

施引文献

资源附件(0)