基于VIS/NIR高光谱技术的灵武长枣SSC含量的无损检测研究

Application of Vis/NIR Hyperspectral Imaging Technology in Non-Destructive Measurement of Soluble Solid Content in Lingwu Jujube

摘要:

利用高光谱技术对灵武长枣可溶性固形物含量（SSC）检测进行研究，为灵武长枣内部品质无损检测提供科学方法。以灵武长枣为对象，对光谱数据进行预处理，应用连续投影算法（SPA)和正自适应加权算法（CARS）进行关键波长的选择，通过偏最小二乘回归（PLSR）和主成分回归（PCR）建立预测模型。实验结果表明：采用去趋势（Detrend）预处理算法效果最优，PLSR模型的交叉验证相关系数（Rcv）为0.809，交叉验证均方根误差（RMSECV）为1.331；通过SPA算法和CARS算法能够有效地对光谱数据进行降维处理，对SPA优选的8个和CARS优选的21个特征变量分别用PLSR和PCR建立预测模型，CARS-PLSR建模效果最优，其相关系数（Rp）为0.864，预测均方根误差（RMSEP）为1.174；研究结果表明基于高光谱成像技术采集的灵武长枣漫反射光谱进行SSC无损检测具有可行性。

Abstract:

Hyperspectral imaging technology was adopted to measure the soluble solid content (SSC) of Lingwu jujube fruits and provided a scientific method for the non-destructive measurement of their interior quality. The diffuse reflectance spectra of Lingwu jujube fruits were preprocessed, the successive projections algorithm (SPA) and competitive adaptive reweighed sampling (CARS) were used to select the characteristic wavelengths, and the partial least squares regression (PLSR) model and principal component regression (PCR) were employed to build the predictive model for the SSC of Lingwu jujube. The results indicated that the detrend pretreatment method provided the optimum performance, the correlation coefficient of cross calibration (Rcv) of the established PLSR model was 0.809, and the root mean square error of cross validation (RMSECV) was 1.331. The SPA and CARS were effective in the dimensionality reduction of spectral data. Based on the eight and 21 characteristic variables selected by SPA, predictive models were established using PLSR and PCR, respectively. The optimal prediction performance was presented by the CARS-PLSR model, whose correlation coefficient of prediction (Rp) and root mean square error of prediction (RMSEP) were 0.864 and 1.174, respectively. The results indicate that non-destructive measurement of the SSC of Lingwu jujube using hyperspectral imaging technology is feasible.