苹果可溶性固形物含量的检测方法——基于可见光近红外光谱技术

Prediction Model for Soluble Solids Content of Apple——Based on Visible/Near Infrared Spectroscopy Technology

为了快速检测苹果的可溶性固形物(SSC)含量,采用可见光近红外光谱技术,结合主成分分析(PCA)和BP神经网络技术,来建立苹果SSC的预测模型。获取苹果样本在345～1039nm波段的漫反射光谱,采用DPS数据处理系统对其进行主成分分析,并提取出累计可信度大于95%的5个新主成分。建立一个3层的BP神经网络模型,并将这5个新的主成分作为BP神经网络模型的输入量,其结果是98%以上预测样本的预测相对误差在5%以下。该研究表明,采用近红外光谱技术来建立苹果可溶性固形物的预测模型是可行的。

To predict the soluble solid content （SSC） of apples quickly, this paper reports on a new prediction model for apple＇s SSC by means of visible/near infrared spectroscopy （Vis/NIRS） （345 - 1039 ）. The reflectance spectra were acquired by using the spectroscopy. Principal component analysis was used to analyze the characteristics of the spectra. Five new principal components were conf/rmed to describe the whole body of spectra when the total reliabilities arrived at 95%. Then the five new principal components were applied as the input of back propagation （BP） neural network with one hidden layer. The values of SSC were applied as the output of BP neural network. The results showed that the relative error of over 98% prediction samples was under 5% and this prediction model was reliable and practicable.