Identification of Rice Varieties Based on Terahertz Time Domain Spectroscopy
-
摘要: 为实现大米品种的准确鉴别，提出一种基于太赫兹时域光谱（Terahertz Time-Domain Spectroscopy, THz-TDS）技术的大米品种识别方法。利用标准差（Standard Deviation, SD）和区间偏最小二乘（Interval Partial Least Square, iPLS）选取0.53~1.21 THz波段的吸收光谱信息作为分类模型的输入数据，再采用决策树模型（Decision Tree, DT）对四种大米吸收光谱进行分类识别，并在模型训练过程中结合网格搜索算法寻找模型最优参数。为增加实验对比度，分别使用逻辑回归模型和支持向量机模型进行对比实验，其模型分类准确率分别为80.75%和88.75%。实验结果表明，太赫兹时域光谱技术结合SD、iPLS和DT方法可以实现大米品种的准确识别，准确率可达95%，为农产品品种识别提供了一种新的鉴别方法。Abstract: In order to achieve accurate identification of rice varieties, a rice variety identification method based on terahertz time-domain spectroscopy (THz-TDS) technology was proposed. To select the terahertz band by Interval partial least square (iPLS) and standard deviation (SD), which determined the absorption spectrum data of the 0.53~1.21 THz band as the input data of the classification model. Then to use the decision tree (DT) identified absorption spectra of four kinds of rice and the model parameters was obtained by combining with the grid search algorithm. In order to increase the experimental contrast, logistic regression models and support vector machine models were used separately for comparative experiments, and the model classification accuracy was 80.75% and 88.75%, respectively. Experimental results showed that the terahertz time-domain spectroscopy technology combined with SD, iPLS and DT methods can realize the accurate identification of rice varieties with an accuracy rate of up to 95%, providing a new identification method for the identification of agricultural varieties.
-
图 2 太赫兹时域光谱系统透射式（a）和反射式（b）
Figure 2. Terahertz time domain spectroscopy system transmission type (a) and reflection type (b)
图 3 时域光谱图（a）和频域光谱图（b）
Figure 3. Time domain spectroscopy (a) and frequency domain spectroscopy (b)
图 6 四种样品预处理后不同波段标准差
Figure 6. Standard deviation of different bands after pretreatment of four samples
表 1 不同样品光谱角度
Table 1. Different sample spectral angles
红米与珍珠糯米 红米与黑米 红米与富硒大米 原始光谱角度 0.077 0.093 0.060 标准化预处理后 2.513 2.652 1.182 
下载: 导出CSV
表 2 四种样品预处理后不同波段标准差
Table 2. Standard deviation of different bands after pretreatment of four samples
太赫兹波
段(THz)河南红
米样品珍珠糯
米样品黑米样品 富硒大
米样品标准差
平均值0~0.53 0.65 0.53 0.36 0.12 0.42 0.53~1.21 0.06 0.07 0.08 0.06 0.07 1.21~1.74 0.40 0.19 0.45 0.18 0.31 1.74~2.31 0.29 0.19 0.25 0.20 0.23 2.31~2.71 0.28 0.28 0.27 0.25 0.27 2.71~3.40 0.25 0.30 0.30 0.23 0.27 3.40~4.00 0.24 0.25 0.28 0.20 0.24 0~4.00 0.59 0.38 0.45 0.21 0.41
下载: 导出CSV
表 3 iPLS不同分割区间下最优区间列表
Table 3. List of optimal intervals under different segmentation intervals of interval partial least square
区间个数 最佳区间 最优因子 均方根误差 对应太赫兹波段（THz） 2 2 3 0.84 0~1.98 3 1 9 0.50 0~1.32 4 1 8 0.62 2.03~3.00 5 1 8 0.60 1.61~2.40 6 4 15 0.57 0.64~1.32 7 2 21 0.55 0.55~1.12 8 2 5 0.52 0.50~0.99 10 2 5 0.57 0.81~1.19 15 3 9 0.54 0.53~0.81 22 4 3 0.60 0.92~1.08
下载: 导出CSV
表 4 决策树分类准确率
Table 4. Classification accuracy of decision tree
模型 测试集准确率（%） 河南红米 珍珠糯米 黑米 富硒大米 平均准确率 DT 100 80 100 100 95
下载: 导出CSV
表 5 不同核函数分类准确率
Table 5. Classification accuracy of different kernel functions
不同核函数 测试集准确率（%） 河南红米 珍珠糯米 黑米 富硒大米 平均准确率 RBF 100 60 85 25 67.5 linear 100 100 80 75 88.75 Sigmoid 100 40 80 25 61.25
下载: 导出CSV
表 6 不同模型的分类准确率
Table 6. Classification accuracy of different models
模型 平均准确率（%） LC 80.75 SVM 88.75 SD-iPLS-DT 95
下载: 导出CSV
-
[1] 周广春, 孟维韧, 全东兴, 等. 吉林省第八届优质食味水稻品种鉴评报告[J]. 东北农业科学,2018,43(6):1−4. [ZHOU G C, MENG W R, QUAN D X, et al. A report on the eighth session of evaluation of high eating quality rice varieties in Jilin Province[J]. Journal of Northeast Agricultural Sciences,2018,43(6):1−4. [2] 张玉荣, 周显青, 杨兰兰. 大米食味品质评价方法的研究现状与展望[J]. 中国粮油学报,2009,24(8):155−160. [ZHANG Y R, ZHOU X Q, YANG L L. Present situation and expectation on methods for taste evaluation of rice[J]. Journal of the Chinese Cereals and Oils Association,2009,24(8):155−160. [3] CAMPOS I, ALCANIZ M, AGUADO D, et al. A voltammetric electronic tongue as tool for water quality monitoring in wastewater treatment plants[J]. Water Research,2012,46(8):2605−2614. doi: 10.1016/j.watres.2012.02.029 [4] 刘敏, 谭书明, 张洪礼, 刘芳宏. 不同品种大米口感品质分析[J]. 食品科学,2018,39(15):88−92. [LIU M, TAN S M, ZHANG H L, et al. Taste quality of different rice varieties[J]. Food Science,2018,39(15):88−92. doi: 10.7506/spkx1002-6630-201815013 [5] 刘晓欢, 刘翠玲, 孙晓荣, 等. 基于傅里叶红外光谱技术的大米产地溯源快速判别方法研究[J]. 食品科技,2021,46(4):244−249. [LIU X H, LIU C L, SUN X R, et al. Study on fast identification method of rice origin traceability based on fourier transform infrared spectroscopy technology[J]. Food Science and Technology,2021,46(4):244−249. [6] 赵凡, 闫昭如, 薛建新, 等. 高光谱无损识别野生和种植黑枸杞[J]. 光谱学与光谱分析,2021,41(1):201−205. [ZHAO F, YAN Z R, XUE J X, et al. Identification of wild black and cultivated Goji berries by hyperspectral lmage[J]. Spectroscopy and Spectral Analysis,2021,41(1):201−205. [7] 惠延波, 潘印卿, 王莉, 等. 基于电子舌技术的不同品种大米区分识别研究[J]. 科学技术与工程,2014,14(33):250−254, 259. [HUI Y B, PAN Y Q, WANG L, et al. Discrimination of different varieties of rice base on electronic tougue technique[J]. Science Technology and Engineering,2014,14(33):250−254, 259. doi: 10.3969/j.issn.1671-1815.2014.33.047 [8] 王靖会, 程娇娇, 刘洋, 等. 基于高光谱成像技术鉴别大米品种[J]. 中国农业科技导报,2021,23(9):121−128. [WANG J H, CHEN J J, LIU Y, et al. Identification of rice variety based on hyperspectral imaging technology[J]. Journal of Agricultural Science and Technology,2021,23(9):121−128. [9] 王朝辉, 高地, 赖瀚清, 等. 近红外光谱法对松原不同品种大米的确证探究[J]. 食品研究与开发,2020,41(4):146−155. [WANG C H, GAO D, LAI H Q, et al. Confirmation of different rice varieties in Songyuan by near infrared spectroscopy[J]. Food Research and Development,2020,41(4):146−155. doi: 10.12161/j.issn.1005-6521.2020.04.025 [10] 刘亚超, 李永玉, 彭彦昆, 等. 近红外二维相关光谱的掺和大米判别[J]. 光谱学与光谱分析,2020,40(5):1559−1564. [LIU Y C, LI Y Y, PENG Y K, et al. Application of two-dimensional correlation spectra in the identification of adulterated rice[J]. Spectroscopy and Spectral Analysis,2020,40(5):1559−1564. [11] 黄星奕, 张浩玉, 赵杰文. 电子舌技术在食品领域应用研究进展[J]. 食品科技,2007(7):20−24. [HUANG X Y, ZHANG H Y, ZHAO J W. Research and application of electronic tongue technology in food industry[J]. Food Science and Technology,2007(7):20−24. doi: 10.3969/j.issn.1005-9989.2007.07.006 [12] 云雯, 吴彦蕾, 杨彦, 等. 电子舌在食品安全检测领域的研究进展[J]. 中国调味品,2014,39(7):133−137. [YUN W, WU Y L, YANG Y, et al. Research progress of application of electronic tongue technique in food safety testing[J]. China Condiment,2014,39(7):133−137. doi: 10.3969/j.issn.1000-9973.2014.07.034 [13] 曾凤仙, 周道志. 红外光谱技术及其在食品检测中的应用[J]. 现代食品,2018(16):131−133. [ZENG F X, ZHOU D Z. Infrared spectroscopy and its application in food detection[J]. Modern Food,2018(16):131−133. [14] 陈超, 柳琦, 李钒, 等. 红外光谱技术在食品安全检测中的研究与应用[J]. 食品研究与开发,2019,40(14):219−224. [CHEN C, LIU Q, LI F, et al. Research and application of infrared spectroscopy technology in food safety testing[J]. Food Research and Development,2019,40(14):219−224. [15] 曾子琦, 蒋立文, 刘霞, 等. 近红外光谱无损检测在食用油脂分析中的应用研究进展[J]. 中国油脂,2018,43(8):137−142. [ZENG Z Q, JING L W, LIU X, et al. Advance in application of near infrared spectroscopy in nondestructive analysis of edible oil[J]. China Oils and Fats,2018,43(8):137−142. doi: 10.3969/j.issn.1003-7969.2018.08.029 [16] 滕安国, 高峰, 夏新成, 等. 高光谱技术在农业中的应用研究进展[J]. 江苏农业科学,2009(3):8−11. [TENG A G, GAO F, XIA X C, et al. Research progress on the application of hyperspectral technology in agriculture[J]. Jiangsu Agricultural Sciences,2009(3):8−11. doi: 10.3969/j.issn.1002-1302.2009.03.003 [17] MANJAPPA M, SINGH R. Materials for terahertz optical science and technology[J]. Advanced Optical Materials,2020,8(3):1901984. doi: 10.1002/adom.201901984 [18] LI T, ZHANG L, HE J A, et al. Terahertz time-domain spectroscopy for identification of hazardous substances in mail[J]. Spectroscopy and Spectral Analysis,2019,39(12):3724−3730. [19] LIU X H, ZHAO H W, LIU G F, et al. Application of terahertz technology in pharmaceutical setting[J]. Progress in Chemistry,2010,22(11):2191−2198. [20] OUYANG A G, ZHANG Y L, LI B, et al. Study on the detection method of alum content in sweet potato starch by terahertz spectroscopy[J]. Spectroscopy and Spectral Analysis,2020,40(3):727−732. [21] 刘燕德, 杜秀洋, 李斌, 等. 太赫兹时域光谱技术对紫米掺假的检测研究[J]. 光谱学与光谱分析,2020,40(8):2382−2387. [LIU Y D, DU X Y, LI B, et al. Detection of purple rice adulteration by terahertz time domain spectroscopy[J]. Spectroscopy and Spectral Analysis,2020,40(8):2382−2387. [22] JU X G, ZHANG Y, LIAN F Y, et al. Quick test for transgenic components in rice using terahertz spectra[J]. Applied Spectroscopy,2019,73(2):171−181. doi: 10.1177/0003702818812085 [23] 孙彤, 张卓勇, 相玉红, 等. 利用Far-IR和THz-TDS光谱法测定大米中西维因[J]. 光谱学与光谱分析,2016,36(2):541−544. [SUN T, ZHANG Z Y, XIANG Y H, et al. Determination of carbaryl in rice by using Far-IR and THz-TDS techniques[J]. Spectroscopy and Spectral Analysis,2016,36(2):541−544. [24] JIANG Y Y, GE H Y, ZHANG Y. Quantitative determination of maltose concentration in wheat by using terahertz imaging[J]. Spectroscopy and Spectral Analysis,2018,38(10):3017−3022. [25] 耿响, 张泽栋, 江龙发, 等. 基于近红外光谱技术的油茶籽粕中灰分含量快速检测方法[J]. 红外,2022,43(1):43−48. [GENG X, ZHANG Z D, JIANG L F, et al. A fast detection method of ash content in camellia seed meal based on near infrared spectroscopy technology[J]. Infrared,2022,43(1):43−48. doi: 10.3969/j.issn.1672-8785.2022.01.007 [26] 张优优, 陈伟豪, 唐志敏, 等. 区间偏最小二乘结合差分进化算法应用于鱼粉近红外光谱波长筛选[J]. 分析测试学报,2020,39(11):1392−1397. [ZHANG Y Y, CHEN W H, TANG Z M, et al. Application of interval partial least squares with differential evolution algorithm in wavelength selection of near infrared spectroscopy for fishmeal[J]. Journal of Instrumental Analysis,2020,39(11):1392−1397. doi: 10.3969/j.issn.1004-4957.2020.11.012 [27] 张坤. 决策树CART在高校学生体质测试数据分析中的应用研究[J]. 四川体育科学,2022,41(1):41−45. [ZHANG K. Application of decision tree CART in data analysis of physical fitness test of college students[J]. Sichuan Sports Science,2022,41(1):41−45. doi: 10.13932/j.cnki.sctykx.2022.01.10 [28] DORNEY T D, BARANIUK R G, MITTLEMAN D M. Material parameter estimation with terahertz time-domain spectroscopy[J]. Journal of the Optical Society of America,2001,18(7):1562−1571. doi: 10.1364/JOSAA.18.001562 [29] DUVILLARET L, GARET F, COUTAZ J L. Highly precise determination of optical constants and sample thickness in terahertz time-domain spectroscopy[J]. Applied Optics,1999,38(2):409−415. doi: 10.1364/AO.38.000409 [30] 张昊丹, 孙孝林, 王晓晴, 等. 可见-近红外光谱原位估测土壤有机质含量中测试点和采样位置导致的误差分析[J]. 光谱学与光谱分析,2020,40(11):3499−3507. [ZHANG H D, SUN X L, WANG X Q, et al. Analyzing errors due to measurement positions and sampling locations for in situ measurements of soil organic matter using vis-nir spectroscopy[J]. Spectroscopy and Spectral Analysis,2020,40(11):3499−3507. [31] 熊芩, 张若秋, 李辉, 等. 最小角回归算法(LAR)结合采样误差分布分析(SEPA)建立稳健的近红外光谱分析模型[J]. 分析测试学报,2018,37(7):778−783. [XIONG Q, ZHANG R Q, LI H, et al. A robust near infrared modeling by least angel regression and sampling error profile analysis[J]. Journal of Instrumental Analysis,2018,37(7):778−783. doi: 10.3969/j.issn.1004-4957.2018.07.004 [32] 王世芳, 韩平, 崔广禄, 等. SPXY算法的西瓜可溶性固形物近红外光谱检测[J]. 光谱学与光谱分析,2019,39(3):738−742. [WANG S F, HAN P, CUI G L, et al. The NIR detection research of soluble solid content in watermelon based on SPXY algorithm[J]. Spectroscopy and Spectral Analysis,2019,39(3):738−742. [33] 郭慧娴, 朱思祁, 黎远鹏等. 基于iPLS和SiPLS算法的人体血清胆红素含量的可见-近红外光谱建模[J]. 光电子激光,2016,27(10):1136−1144. [GUO H X, ZHU S Q, LI Y P, et al. Visible-near infrared spectroscopy modeling on the contents of serum bilirubin based on iPLS and SiPLS[J]. Journal of Optoelectronics Laser,2016,27(10):1136−1144. [34] GUO J, DENG H, CHEN L J, et al. A reliable method for identification of antibiotics by terahertz spectroscopy and svm[J/OL]. Journal of Spectroscopy, 2020. https://doi.org/10.1155/2020/8811467 [35] 王欣卉, 宋雪健. 近红外光谱技术对稻花香大米品牌的保护研究[J]. 粮食科技与经济,2020,45(11):97−99. [WANG X H, SONG X J. Study on the protection of Daohuaxiang rice brand by near infrared spectroscopy[J]. Grain Science and Technology and Economy,2020,45(11):97−99. doi: 10.16465/j.gste.cn431252ts.20201128 [36] 翁士状, 唐佩佩, 张雪艳, 等. 高光谱成像的图谱特征与卷积神经网络的名优大米无损鉴别[J]. 光谱学与光谱分析,2020,40(9):2826−2833. [WENG S Z, TANG P P, ZHANG X Y, et al. Non-destructive identification method of famous rice based on image and spectral features of hyperspectral imaging with convolutional neural network[J]. Spectroscopy and Spectral Analysis,2020,40(9):2826−2833. [37] 沙敏, 李良翠, 黄家乐, 等. 拉曼光谱数据处理方式对大米产地鉴别模型的影响[J]. 中国食品学报,2021,21(5):369−376. [SHA M, LI L C, HUANG J L, et al. Effect of data processing method on identification of rice from different geographical origins by raman spectroscopy[J]. Journal of Chinese Institute of Food Science and Technology,2021,21(5):369−376. doi: 10.16429/j.1009-7848.2021.05.044 -
点击查看大图
图(7) / 表(6)
计量
- 文章访问数: 87
- HTML全文浏览量: 25
- PDF下载量: 9
- 被引次数: 0
