Authors

Abstract

Mold contamination of stored maize can cause significant economic losses, and it is crucial to effectively classify maize kernels without destroying their original structure. But existing studies have found it difficult to distinguish moldy maize. In this paper, a method for non-destructive detection of mold in maize using near-infrared spectral fingerprinting is proposed. The spectral raw data are initially acquired using a handheld near-infrared spectrometer. To enhance the signal quality, preprocessing is conducted, and a classification model is developed for full-band spectral data. In order to further optimize the model and enhance the classification accuracy, the feature wavelengths were extracted from the spectral data with effective preprocessing techniques in the full-band model. Finally, the maize kernel mold classification model is constructed. The classification accuracy of SG+SNV-SVM-ISFLA model can reach up to 97.22%, and the accuracy for the identification of asymptomatic moldy maize is 96.30%, which can realize the accurate grading of moldy accurate classification of maize and can well distinguish asymptomatic moldy maize. This work may significantly control the spread of molds in the food industry while improving storage economics and safety.

Abstract in Chinese

仓储玉米受到霉菌污染会造成重大经济损失，因此在不破坏玉米原有结构的情况下对玉米进行有效分类至关重要。但现有研究发现，轻微霉变玉米难以区分。本文提出了一种利用近红外光谱指纹的玉米霉变无损检测的方法。最初使用手持式近红外光谱仪获取光谱原始数据。为了提高信号质量，进行了预处理，并为全波段光谱数据建立分类模型。为了进一步优化模型并提高分类准确率，对预处理后的数据进行特征提取。最后，构建了玉米霉变分类模型。结果表明SG+SNV-ISFLA-SVM模型的分类准确率高达97.22%，对无症状霉变玉米的识别准确率为96.30%，可实现对玉米霉变的准确分级，并能很好地区分无症状霉变玉米。这项工作可大大控制霉菌在食品工业中的传播，同时提高贮藏的经济性和安全性。