基于激光诱导荧光高光谱技术无损检测脐橙表面敌敌畏残留  

Nondestructive Detecting Dichlorvos Residue on Navel Orange Surface Basedon Laser Induced Fluorescence Hyperspectral Image Technique

作　　者：薛龙[1,2] 庄宏[1] 黎静[2] 刘木华[2,3] 王晓[2] 罗春生[2] 

机构地区：[1]华东交通大学机电工程学院,南昌市330013 [2]江西农业大学工学院生物光电实验室,南昌市330045 [3]无损检测技术教育部重点实验室,南昌市330029

出　　处：《中国农机化》2012年第1期189-193,共5页Chinese Agricul Tural Mechanization

基　　金：国家自然科学基金资助项目(30760101);新世纪优秀人才支持计划资助项目(NCET-09-0168);江西省科技支撑计划项目(2009BNB05705);江西省青年科学家(井岗之星)培养

摘　　要：常规化学方法检测农药残留不仅对样品具有破坏性,而且费时费力。本文以激光诱导荧光结合高光谱图像技术为手段,对脐橙表面的敌敌畏农药残留进行光谱无损检测;实验方法是在脐橙表面,喷施用自来水配制的不同浓度的敌敌畏农药溶液,在实验室条件下风干后,采集激光诱导荧光高光谱图像,再用气相色谱法检测脐橙表面的农药残留量,应用偏最小二乘（Partial least squares,PLS）方法建立农药残留的预测模型,并找出最佳光谱区间,然后应用支持向量机（Support vectormachine,SVM）方法在最佳光谱区间的基础上建立农药残留的预测模型;所建模型结果其预测集样品的农药残留量实测值（0.4862~10.3791mg/kg）和预测值之间的相关系数为0.8101;实验结果说明,以激光诱导荧光结合高光谱技术为手段的无损检测技术,在检测脐橙农药残留方面是有可行性的。Pesticide residue detection using conventional chemical methods is destructive to samples,and it is a time and labor consuming work.In this study,laser induced fluorescence combined with hyperspectral image technique was used to nondestructive detect dichlorvos residue on navel orange surface.Navel oranges were sprayed with different concentration dichlorvos solutions.After the samples have been air-dried,the laser induced fluorescence combined with hyperspectral images were collected.The actual amounts of dichlorvos residue on navel oranges were measured using gas chromatography（GC）.The prediction model was built using the partial least squares（PLS） method and the optimal spectral regions were extracted.Based on the optimal spectral regions,support vector machine（SVM） was used to build prediction model.The correlation coefficient（r） of the measured values（0.4862~10.3791mg / kg） and predicted values was 0.8101.The experimental result shows that it is feasible to nondestructive detect dichlorvos residue using laser induced fluorescence combined with hyperspectral image technique.

关 键 词：无损检测 敌敌畏残留 激光诱导荧光 高光谱图像 脐橙 

分 类 号：O433.5[机械工程—光学工程]