人工神经网络建模与高光谱数据预处理相结合用于鉴别阿胶掺假

摘要：研究了高光谱成像与光谱预处理和神经网络技术相结合，以实现阿胶（（ CCA，阿胶）掺假鉴别。CCA 掺入猪皮明 胶（（ PSG），掺假比例为 5-95%（（ w/w），每次增加 5%。使用三种方法对原始光谱进行预处理，即乘性散射校正（（ MSC）、 Savitzky-Golay（（ SG）平滑处理以及 MSC 和 SG 的组合（（ MSC-SG）。使用 SPA 选择特征波长（（ CW）以降低高维。提 取 CW 的颜色和纹理特征作为预测模型的输入。采用两种人工神经网络（ANN）结合三种光谱预处理方法建立预测模 型，其中广义回归神经网络（GRNN）结合 MSC-SG 预处理方法的预测模型表现出良好的效果，正确分类率达 92.5%。 结果表明，预处理方法、高光谱成像特征和ANN建模相结合对CCA掺假鉴别具有很大的潜力和可行性。

期刊：Journal of Food Quality

DOI：https://DOI.org/10.1155/2018/3487985

设备：GaiaField

Integration of Artificial Neural Network Modeling and Hyperspectral Data Preprocessing for Discrimination of Colla Corii Asini Adulteration

Abstract：The studyofhyperspectral imaging in tandemwithspectral preprocessing and neural network techniques was conducted to realize Colla Corii Asini (CCA, E’jiao) adulteration discrimination. CCA was adulterated with pig skin gelatin (PSG) in the range of 5–95% (w/w)at5%increments.Threemethodswereusedtopretreattheoriginalspectra,whicharemultiplicativescattercorrection(MSC), Savitzky-Golay (SG) smoothing, and the combination of MSC and SG (MSC-SG). SPA was employed to select the characteristic wavelengths (CWs)toreduce thehighdimension.Colourandtexture features of CWs wereextracted as input of prediction model. Two kinds of artificial neural network (ANN) with three spectral preprocessing Methods were applied to establish the prediction models.Thepredictionmodelofgeneralizedregressionneuralnetwork(GRNN)intandemwiththeMSC-SGpreprocessedmethod presented satisfactory performance with the correct classification rate value of 92.5%. The results illustrated that the integration of preprocessing methods, hyperspectral imaging features, and ANN modeling had a great potential and feasibility for CCA adulteration discrimination.