寒富苹果品质无损检测光谱信息在线分析

Online Analysis on Hanfu Apple Quality Spectral Information Based on NDT

苹果深加工和商品化等问题已成为制约苹果产业发展的瓶颈问题,针对苹果品质的无损检测具有现实意义。应用USB2000+可见/近红外光纤光谱仪对寒富苹果内部品质信息在线检测的可行性进行研究,分析在检测距离(37±1)mm,果盘运行速度0.2m·s-1,检测时间间隔为1s的前提下,不同的光谱预处理方法对苹果各品质模型的影响,并应用偏最小二乘法(PLS)对寒富苹果的可溶性固形物、总酸度及硬度做出相应的预测模型。结果表明:经二阶导数结合标准正态变量变化(SD+SNV)预处理方法建立的可溶性固形物预测模型最优,模型的测定系数(R2)0.9743,校验标准差(RMSEC)0.1503,预测标准差(RMSEP)0.1636。模型检验结果表明:残差分布小于4%,相对误差小于3%,说明模型预测能力可靠,应用USB2000+光纤光谱仪在线检测寒富苹果内部品质可行,满足苹果品质检测要求。

Further processing and commercialization of apple have increasingly become the bottleneck of the development of apple industry. It has practical significances to improve its utilization value and to promote the industry development. The purpose of this study is to investigate the feasibility of Hanfu apple internal quality information on-line detection using USB2000 +visible/near infrared optical fiber spectrometer. The tests were performed at the detection distance of(37 ±1)mm,compote running speed of 0.2 m·s-1and test interval of 1s. Some spectral preprocessing methods including multiple scatter correction, derivative and both of them combined were compared. At the same time, the corresponding prediction mathematical model for the soluble solids, total acidity and hardness of hanfu apple was established by the means of partial least squares(PLS). The results showed that soluble solids prediction model got by the second derivative combined with standard normal variables change(SD + SNV) gave the best performance, in which the determination coefficient(R2) was 0.9743, the checking standard deviation(RMSEC) was 0.1503 and the predicting standard deviation(RMSEP) was 0.1636. Predictive ability of the model proves convincing and reliable by the residual distribution and the relative error were less than 4% and 3% respectively,which indicated that the application visible/near-infrared spectroscopy on the on-line detection for hanfu apple internal quality is feasible and can meet the requirements of apple quality detection.