基于高光谱技术的黑果腺肋花楸成熟度判别及多酚含量检测模型构建

Establishment of Models for Maturity Discrimination and Polyphenol Content Determination of Aronia melanocarpa Using the Hyperspectral Technology

为了无损检测黑果腺肋花楸(Aronia melanocarpa,简称黑果)果实成熟度和多酚含量,本研究构建了基于高光谱成像技术的黑果成熟度判别模型以及多酚含量检测模型。采用高光谱成像技术采集不同成熟度的富康源1号黑果图像信息,福林酚法测定其多酚含量。通过蒙特卡洛法剔除异常值;滑动平均、中值滤波、归一化、基线校准、多元散射校正、消除趋势和标准正态变量变换对原始图像信息进行预处理;光谱-理化值共生距法进行样本划分;竞争性自适应重加权算法和无信息变量消除法提取特征波长,分别建立偏最小二乘模型(PLS)和支持向量机(SVM)模型并进行比较。结果表明,本研究建立的判别模型中效果最好的模型为经多元散射校正预处理后的UVE-SVM模型,综合识别率94.62%,R_(c)^(2)=0.9712,根据该模型判别的准确度为100%。多酚含量检测效果最好的模型为中值滤波预处理后的CARSSVM模型,R_(c)^(2)=0.8331。此外,本研究还证明了黑果多酚含量的可视化是可行的。本研究为高光谱成像技术在浆果领域的应用提供了理论基础。

Models for maturity discrimination and polyphenol content determination of Aronia melanocarpa(AM)were established to nondestructively determinethe maturityand polyphenol content of AM based on the hyperspectral imaging(HSI)technology.The HSI technology was adopted to collect the image information of Fukangyuan No.1 AM in different maturity levels,and the Folin-Ciocalteu colorimetric method was employed to determine the polyphenol content.The Monte Carlo method was used to eliminate the outliers.The original image information was preprocessed with the following procedures:Moving average,median filter,normalize,baseline calibration,multiple scattering correction,detrending,and standard normal variate.Meanwhile,the sample set partitioning based on joint x-y distance method was applied to divide the samples.The competitive adaptive reweighting sampling and uninformative variable elimination method were selected to extract the feature wavelengths,based on which the partial least squares model(PLS)and support vector machine(SVM)model were established and compared.The results showed that the UVE-SVM model after multiple scattering correction showed the best performance among all established discriminant models,with the comprehensive recognition rate of 94.62%,R_(c)^(2) of 0.9712,and accuracy of 100%.The CARS-SVM model after median filter exhibited the best efficiency in detection of polyphenol content,with the R_(c)^(2) of 0.8331.In addition,this work proved that the polyphenol content in AM was visual.This work provided a theoretical basis for the application of HSI technology in berry industry.