论文部分内容阅读
为了能够客观、快速、无损、便捷地检测花生仁霉变和出芽,研究了一种基于傅里叶变换近红外光谱技术和K最近邻(KNN)模式识别方法的霉变和出芽花生识别方法。依据花生的感官特征和前人研究经验,将花生分为正常、轻度霉变、重度霉变和发芽四类,采用傅里叶变换近红外光谱仪的积分球漫反射方法采集花生光谱(波段4 000~10 000 cm-1)。利用二阶导数算法进行光谱预处理,建立联合区间偏最小二乘(Si-PLS)识别模型,并得到特征光谱区间。然后在特征光谱区间的基础上运用主成分分析进行数据空间降维,最后建立KNN识别模型。KNN模型训练集与预测集识别率均为98.84%,表明应用近红外光谱技术和KNN法检测霉变和出芽花生效果良好,具有可行性。
In order to objectively, rapidly, non-destructively and conveniently detect peanut almonds and sprouts, a method of recognizing moldy and sprouted peanuts based on Fourier transform near-infrared spectroscopy and K nearest neighbor (KNN) pattern recognition was studied. According to the sensory characteristics of peanut and previous research experience, the peanut was divided into four categories: normal, mild mildew, severe mildew and germination. The pears were collected using the Fourier transform near-infrared spectroscopy diffuse reflectance spectroscopy (band 4 000 to 10,000 cm-1). The second-order derivative algorithm is used to preprocess the spectrum, and a joint interval partial least-squares (Si-PLS) identification model is established, and the characteristic spectral interval is obtained. Then based on the characteristic spectral interval, principal component analysis is used to reduce the dimension of data space. Finally, a KNN identification model is established. The recognition rate of KNN model training set and prediction set are both 98.84%, which shows that it is feasible to detect mildew and budding peanut by near infrared spectroscopy and KNN method.