基于光谱指数的蜜橘成熟度评价模型研究

本文探索了基于光谱指数的蜜橘成熟度快速无损评价方法及模型。以2016年9~11月份6个不同采收期的300个蜜橘作为实验样品,采集重量、横纵径、叶绿素、色差、可溶性固形物(SSC)、酸度(TA)、近红外光谱等数据。通过对比分析上述各参数的平均值和偏差,筛选出叶绿素、叶绿素/SSC、叶绿素/固酸比作为蜜橘成熟度评价指标。利用光谱变异系数分析光谱的特征,筛选出649、724、672、1 100 nm 4个特征波长,通过特征波长线性组合方法以及相关性分析,得出最佳光谱指数。接着,以225个样品为建模集、75个样品为预测集,在成熟度评价指标与光谱指数间进行多元线性回归(MLR)分析。对比发现,以叶绿素为成熟度评价指标的评价模型的预测结果最准确,建模和预测相关系数分别达到0.98和0.96,建模均方根误差(RMSEC)和预测均方根误差(RMSEP)分别为0.49和0.59,建模和预测偏差分别为-6.1×10-8和-0.014。实验结果表明,利用光谱指数能便捷、准确地评定蜜橘成熟度,为后续开发低成本测量成熟度的仪器提供了理论依据。

Maturity evaluation model of tangerine based on spectral index

Based on the spectral analysis, we explored a rapid nondestructive evaluation method for maturity of tangerine and estabilizhed a evaluation model. The data of weight, transverse and longitudinal diameter, chlorophyll, color difference, soluble solids(SSC), acidity(TA) and near infrared spectra were collected from 300 tangerines as experimental samples of 6 different picking time from September to November in 2016. The chlorophyll, chlorophyll/SSC and chlorophyll/solid acid ratio were screened out as the evaluation index of maturity of tangerine by comparing and analyzing the average value and deviation of the above parameters. By using spectral coefficients of variation to analyze the characteristics of the spectra, four characteristic wavelengths of 649, 724, 672, and 1 100 nm were selected. The optimal spectral indexes were obtained by linear combination and correlation analysis of these wavelengths. Then, taking 225 samples as model sets and 75 samples as prediction sets, multiple linear regression(MLR) analysis was conducted for both maturity index and spectral index. By comparison, the prediction model based on chlorophyll as the maturity index was the most accurate, and the correlation coefficient between modeling and prediction was 0.98 and 0.96 respectively. The root mean square error of modeling(RMSEC) and root mean square error of prediction(RMSEP) were 0.49 and 0.59, and the modeling and forecast deviations were -6.1×10^-8 and -0.014, respectively. The results showed that the spectral index can be used to conveniently and accurately evaluate the maturity of tangerine, which provided a theoretical basis for the subsequent development of low-cost maturity-measuring instruments.