基于机器学习的NIR光谱柑橘产地鉴别框架

NIR Spectroscopy Identification Framework of Oranges Origin Based on Machine Learning

研究基于机器学习的近红外(Near infrared,NIR)光谱柑橘产地鉴别模型,根据NIR光谱数据的特点,提出了一个完整的产地鉴别通用框架,包括数据预处理、特征选择、模型建立和交叉验证等步骤。在框架下对比多种预处理算法以及多种机器学习算法,基于NIR光谱进行柑橘产地鉴别,得到了较好的识别结果,提高了柑橘产地鉴别的准确性。

This paper studies a machine-learning based near infrared (NIR) spectral model for oranges origin identification. According to the characteristics of NIR spectral data, a complete framework for origin identification is proposed, including data preprocessing, feature selection, model building and cross validation. Then, under this recognition framework, it compares several preprocessing algorithms and a variety of machine learning algorithms. Based on NIR spectroscopy, it also analyzes the oranges origin and obtains better recognition results. This improves the accuracy of oranges origin identification.