| 基于OSC/PLS法识别被不同浓度农药污染水果的初步研究 |
| |
| 引用本文： | 薛龙,黎静,刘木华. 基于OSC/PLS法识别被不同浓度农药污染水果的初步研究[J]. 现代科学仪器, 2010, 0(2): 130-131,137 |
| |
| 作者姓名： | 薛龙 黎静 刘木华 |
| |
| 作者单位： | 1. 江西农业大学工学院,江西南昌,330045;华东交通大学机电工程学院,江西南昌,330013
2. 江西农业大学工学院,江西南昌,330045 |
| |
| 基金项目： | 国家自然科学基金，江西省教育厅科学技术研究项目 |
| |
| 摘 要： | 本文采用正交信号法(orthogonal signal correction,OSC)处理被不同浓度农药污染的脐橙近红外光谱(350～1800nm),在整个光谱范围内应用偏最小二乘法(partialleastsquares,PLS)建立农药污染的预测模型。PLS校正模型采纳的最佳因子数会随着OSC因子的增加而逐渐减少,并且对模型精度影响不明显,因此可以达到简化模型的效果。实验结果表明,当OSC因子数为15时,PLS模型最佳的因子数为3,其预测组脐橙表面农药污染程度的实际类别与预测类别的相关系数R2与预测样本均方根误差RMSEP分别为0.8923和0.3746。
|
| 关 键 词： | 近红外光谱 农药污染 偏最小二乘法（PLS） |
Near Infrared Spectroscopy Recognition and Classification of Different Concentration of Pesticide Contamination on Fruit Surface Based on OSC and PLS |
| |
| Xue Long,Li Jing,Liu Muhua. Near Infrared Spectroscopy Recognition and Classification of Different Concentration of Pesticide Contamination on Fruit Surface Based on OSC and PLS[J]. Modern Scientific Instruments, 2010, 0(2): 130-131,137 |
| |
| Authors: | Xue Long Li Jing Liu Muhua |
| |
| Affiliation: | 1Engineering College,Jiangxi Agricultural University,Nanchang,Jiangxi 330045,China; 2 School of Mechanical and Electronical Engineering,East China JiaoTong University,Nanchang,Jiangxi 330013,China) |
| |
| Abstract: | Orthogonal signal correction（OSC）was used as a method to preprocess the near infrared（NIR）spectra of navel orange,which were sprayed with different concentration of fenvzlerate pesticide.The spectral region was from 3501800nm.The calibration model of recognition and classification of different concentration of pesticide contamination on fruit against navel orange spectra by partial least square（PLS）was established.The study demonstrated that the optimum PLS calibration model was obtained when 15 OSC factors were filtered.The experimental results showed that the coefficient R pred and root mean squared error of prediction（RMSEP）were 0.89230.3746 respectively. |
| |
| Keywords: | Near infrared spectra Pesticide contamination PLS |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
