基于可见-近红外光谱的生鲜紫薯食味品质快速无损检测

目的建立可见-近红外光谱法快速无损检测生鲜紫薯的熟化黏度和甜度,并对其食味品质进行评价。方法以生鲜紫薯为研究对象,用竞争性自适应加权算法(competitive adaptibe reweighted sampling,CARS)筛选出生鲜紫薯预处理后的光谱数据的特征波长,再用偏最小二乘法(partial least square,regression,PLS)建立了生鲜紫薯熟化峰值扭矩和可溶性糖含量的定量预测模型。另外,选取12个生鲜紫薯根据划分分值区间法对紫薯食味品质进行评价。结果生鲜紫薯峰值扭矩模型预测集的相关系数r为0.9195,均方根误差(root mean square error of prediction,RMSEP)为0.0526 Nm;可溶性糖含量模型的预测集的相关系数r为0.9515,RMSEP为0.3100 mg/g。结论基于可见-近红外光谱技术可以对生鲜紫薯食味品质进行初步快速无损评价,为鲜食性紫薯品质评价提供理论参考。

Rapid and nondestructive detection of eating quality of fresh purple sweet potato based on visible-near infrared spectroscopy

Objective To establish a method for the rapid and nondestructive determination of ripening viscosity and sweetness of fresh purple potato by visible-near infrared spectroscopy, and evaluating the taste quality of fresh purple potato. Methods Taking the fresh purple potato powder as the experimental objects, the characteristic wavelength of spectral data pretreated with fresh purple sweet potato was screened by competitive adaptibe reweighted sampling. A quantitative prediction model for peak torque (denoting viscosity) and soluble sugar content of fresh purple potato was established by partial least square regression. In addition, 12 fresh purple potato samples were selected to evaluate their food taste quality by the division score interval method. Results The correlation coefficient of prediction set of the peak torque model of fresh purple potato was 0.9195, and the root mean square error of prediction was 0.0526 Nm; the correlation coefficient of the predicted set of soluble sugar content model was 0.9515, and the root mean square error of prediction was 0.3100 mg/g. Conclusions Preliminary rapid nondestructive detection for the taste quality of fresh purple potato based on visible-near infrared spectroscopy can be achieved, and this method can provide theoretical references for the quality evaluation of fresh purple potato.