基于色敏传感器结合光谱技术的大米储藏期鉴别

Identification of Rice with Different Storage Time Based on Color-sensitive Sensor Array Combined with Visible-near-infrared Spectroscopy

摘要:

应用色敏传感器阵列(CSA)结合可见/近红外(Vis-NIR)光谱检测技术，对大米储藏时间进行鉴别。大米按不同储藏期(0、1、2、4、6个月)分为5组。色敏传感器由氟硼吡咯类色敏材料制成，与大米挥发性气体发生反应后，分别提取色敏材料的光谱数据。光谱数据经 SNV算法预处理后，用Si-PLS算法提取3类光谱数据的最佳光谱区间并合成一个数据集。分别用遗传算法(GA)、无信息变量消除法(UVE)和蚁群算法(ACO)提取光谱变量。并结合主成分分析(PCA)和线性判别分析(LDA)进行模式识别。结果表明，用Si-PLS-UVE提取的光谱变量建立的LDA预测模型正确识别率最高。取主成分数为9时，训练集正确识别率为98%，校正集正确识别率为96%，为大米储藏时间的检测提供了一种可行的方法

关键词:大米;储藏时间;色敏传感器;可见/近红外光谱

Abstract:

Rice is gradually aged during transportation and storage, and its degree of aging is an important factor affecting the quality of rice. A color-sensitive sensor array (CSA) combined with visible-near-infrared spectroscopy system was used to identify rice in different storage periods. Rice samples were stored under constant temperature and humidity conditions and divided into five groups according to different storage periods (0 month, 1 month, 2 months, 4 months and 6 months). CSA was made of three boron-dipyrromethene (BODIPY) dyes to capture the volatile organic compounds. After reacting with volatile organic compounds of rice in different storage periods, the spectral data of the color sensitive materials were separately extracted by the detection system. The optimal spectral range of the three types of spectral data was extracted using the Si-PLS algorithm. After the spectral interval data fusion, the characteristic spectral variables were extracted by genetic algorithm (GA), no information variable elimination method (UVE) and ant colony algorithm (ACO), respectively. Pattern recognition was performed by using principal component analysis (PCA) and linear discriminant analysis (LDA). The results showed that the LDA prediction model established by Si-PLS-UVE extracted spectral variables had the highest recognition rate. When the number of PCs was 9, the correct recognition rate of the calibration set and prediction set was 98% and 96%. The research result provided a viable method for detecting rice storage time.

Key words:rice;storage time;color-sensitive sensor;visible-near-infrared spectroscopy