摘要
自行研制了一套实验室用苹果糖度的近红外光谱在线检测系统,确定了检测光纤与苹果的距离参数,重点探讨了对苹果糖度检测精度影响较大的两个光学参数(入射光纤与检测样品的距离d1、两入射光纤照射光斑的圆心距d2),通过光谱信号稳定性来优化参数条件,优选出了几个参数水平组合。采用多元线性逐步回归(MLR)方法建立糖度预测模型并进一步优选参数组合。研究结果表明本系统在线检测苹果糖度是可行的。
A NIR spectroscopy system was developed as preparation for on-line measure apple＇s sugar content.Some optic parameters of the system were optimized to obtain more effective signal.Optimize the distance between detective fiber was fixed on to ensure sufficient signal.Another two important optic parameters（the distance between the light source and object（d1）,the distance between two incident ray spots on the apple＇s surface（d2） of the system were focused.Spectral signal stability was employed to optimize these optic parameters.Multi linear regression（MLR） method was employed to calibrate prediction model in order to meet the on-line operation.Experimental results showed that the optimal optic parameters were achieved when d1=10 mm and d2=0 mm,and 9 feature wavebands were selected by stepwise MLR method.The study showed it was feasible on-line measure apple＇s sugar content by this system.
出处
《食品科技》
CAS
北大核心
2010年第7期267-270,共4页
Food Science and Technology
基金
国家科技支撑计划项目(2006BAD11A12)
江苏省自然基金重点项目(BK2006707-1)
