基于可见-近红外高光谱成像技术的梨硬度和可溶性固体含量的预测

Soluble Solids Content and Firmness Prediction of Pears Based on Visible-near Infrared Hyperspectral Image

高光谱成像技术可以用在无损检测水果品质方面。可见-近红外波段(400nm至1000nm)的光谱可以用来检测梨的可溶性固体含量(SSC)和硬度。用搭建的可见-近红外光谱范围的高光谱成像系统可以拍摄得到不同种类的梨的光谱图,再通过计算机软件提取光谱图中的有效信息进行分析。用于预测建立模型的两个品种共200个梨样品的光谱图经过连续投影算法(SPA)和多元线性回归算法(MLR)的处理分析可以建立回归模型以及回归方程来有效的预测梨的可溶性固体含量(SSC)和硬度。最终得到硬度模型的相关系数为0.923,SSC模型的相关系数为0.898。

Hyperspectral imaging can be used for the non-destructive quality detection of fruits. Visible-near infrared hyperspectral images( 400-1000nm) are used for the non-destructive detection of soluble solids content( SSC)and firmness of pears. A visible-near infrared imaging spectroscopy system is assembled to acquire scattering images from different varieties of pears. Computer software is developed to extract the information of hyperspectral images.Spectra of 200 pear samples from two varieties are analyzed by successive projections algorithm( SPA) and multiple linear regression( MLR) to predict SSC,firmness of pears. The correlation coefficients are 0. 923 for firmness and 0. 898 for SSC in the prediction set.