可见/近红外光谱的南丰蜜桔可溶性固形物含量定量分析  

ANALYSIS OF SOLUBLE SOLID CONTENT IN NANFENG MANDARIN FRUIT WITH VISIBLE NEAR INFRARED SPECTROSCOPY

作　　者：刘燕德[1] 罗吉[1] 陈兴苗[1] 

机构地区：[1]江西农业大学工学院,江西南昌330045

出　　处：《红外与毫米波学报》2008年第2期119-122,共4页Journal of Infrared and Millimeter Waves

基　　金：国家自然科学基金(60468002;30560064)资助项目

摘　　要：基于可见／近红外漫反射光谱定量分析技术对南丰蜜桔的可溶性固形物含量进行实验研究，采用偏最小二乘法对南丰蜜桔完整果和果肉的可见／近红外光谱进行了分析，并且比较和讨论了不同光谱预处理的建模结果，实验结果表明：在波长范围350—1800nm，一阶微分光谱所建模型效果最佳．其中完整果所建校正模型的预测相关系数为0．825和预测均方根偏差为0．899；果肉所建校正模型的预测相关系数为0．893，预测均方根偏差为0．749。Soluble solids content （SSC） of Nanfeng mandarin fruit was studied by quantitative analysis technique based on visible/near infrared diffuse reflectance spectroscopy （Vis/NIR）. Partial least squares （PLS） regression was carried out to analyze the Vis/NIR of Nanfeng mandarin fruit and its flesh. Calibration results for SSC were compared and discussed by using different spectral pretreatment methods. The result shows that the best calibration models, in the wavelength range of 350 - 1800nm, can be obtained by the first derivative spectrum （ D1 log （l/R） ）. The best prediction results for the mandarin fruit are 0. 825 and 0. 899 for correlation coefficient （ rp ） and root mean square errors of prediction （RMSEP） respective- ly. The best prediction results for its flesh are 0. 893 and 0. 749 for rp and RMSEP respectively.

关 键 词：可见近红外光谱 偏最小二乘法 可溶性固形物含量 南丰蜜桔 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]