-
柑橘是我国栽培面积最大的果树作物,柑橘产业也是南方地区重要的农村经济支柱产业.全国柑橘产业标准化程度不高,导致果实品质优良混杂,采后果实精细分级少,影响柑橘品牌价值和产业经济效益,而进行柑橘果实分级对提高其商品价值和加工附加值具有积极作用[1].目前,我国柑橘产地主要按果实的大小、健康与病果等外表进行果实分级,而分级指标涉及内在品质较少,产地非损伤品质检测和监测技术发展缓慢. W·默科特果实肉质细嫩化渣且多汁,口感酸甜适口且风味浓,是我国南方地区主栽优良晚熟柑橘之一,对当地农业经济发展具有重要意义.
光谱技术具有无损、快速、样品无需前处理等许多独特的优点,利用光谱技术对果实品质进行检测的研究基本已涉及大部分水果[2-6]和果实中可溶性固形物(total soluble solids,TSS)、酸度(titratable acid,TA)、维生素C(vitamin C,Vc)、花青素等品质指标,研究内容主要集中于光谱获取方式、特征波段的选择、建模方法和模型评价等,但采用光谱技术目前对水果品质的检测多为一种模型只能检测一种指标,且模型的通用性较差.因此,开发利用光谱技术以实现果实多指标同时检测、通用性强、精度高的柑橘果实品质检测技术具有重要意义.
本研究以W·默科特果实为对象,分析测定果实中可溶性固形物、可滴定酸和维生素C含量,结合近红外光谱技术,采用间隔偏最小二乘法(interval partial least squares,iPLS)、间隔偏最小二乘法结合连续投影算法(Successive projections algorithm,SPA)和竞争性自适应重加权算法(Competitive adaptive reweighed sampling,CARS)提取特征波长,并采用最小二乘支持向量机回归法(Least square support vector regression,LS-SVR)和偏最小二乘回归法(Partial least squares regression,PLSR)建立果实中3种内在品质指标含量预测模型,以期为柑橘果实内在品质快速无损检测和水果智能分级技术提供参考.
Determination of the Intrinsic Quality of Citrus Variety W·Murcott by Near Infrared Spectroscopy
-
摘要: 探索了近红外光谱技术对W·默科特果实内在多项品质指标进行快速、无损检测的可行性.采集1 000~2 500 nm波段的W·默科特果实的近红外光谱,分别采用间隔偏最小二乘法(iPLS)、竞争性自适应重加权算法(CARS)和间隔偏最小二乘法结合连续投影算法(iPLS-SPA)3种特征波长筛选方法对全波段进行特征波长筛选,利用全波段及3种特征波长筛选方法得到的特征波段对应光谱信息建立了果实可溶性固形物(TSS)、可滴定酸(TA)和维生素C(VC)含量的最小二乘支持向量机回归(LS-SVR)预测模型和偏最小二乘回归(PLSR)预测模型.结果显示,采用CARS筛选的特征波长所对应光谱信息建立的LS-SVR预测模型精度最高,模型对TSS和TA和VC含量的预测集相关系数分别达到0.91,0.85和0.91,且模型对应的预测集均方根误差(RMSEP)分别为0.26,0.03和0.25.说明采用近红外光谱技术结合CARS和LS-SVR可实现对W·默科特果实TSS和TA和VC含量的同时检测.Abstract: The feasibility of fast and nondestructive measurement of main internal quality indexes of the fruit of the citrus variety W·Murcott was investigated using near infrared spectroscopy in this study. Near infrared spectral data of W·murcott fruit were acquired in the spectral region of 1 000-2 500 nm. The interval partial least squares (iPLS) regression, competitive adaptive reweighting algorithm (CARS) and a hybrid variable selection algorithm based on iPLS and successive projections algorithm (iPLS-SPA) were applied for the effective spectral variable selection. Least squares support vector machine regression (LS-SVR) and partial least squares regression (PLSR) were applied to establish a prediction model of the content of soluble solids (TSS), titratable acid (TA) and vitamin C (VC) content in W·murcott fruit. Spectral data of full-spectrum and the effective spectral variable filtered by iPLS, CARS and iPLS-SPA were used in modeling, respectively. The results showed that the LS-SVR prediction model established with the effective spectral variable filtered by CARS best successful, the correlation coefficients of TSS and TA and Vc being 0.91, 0.85 and 0.91, respectively, and the corresponding RMSEP (Root-Mean-Square Error of Prediction) being 0.26, 0.03 and 0.25 respectively. The above results indicated that the simultaneous detection of TSS, TA and VC contents in WIL-1·murcott fruit could be achieved by using near-infrared spectroscopy combined with CARS and LS-SVR.
-
Key words:
- near infrared spectroscopy /
- W·Murcott /
- intrinsic quality /
- content detection .
-
表 1 果实TSS,TA和VC实测值统计
内在品质 样本集 样本数 范围 平均值 标准差 变异系数 TSS/% 校正集 135 9.90~13.20 11.28 0.80 0.07 预测集 45 10.10~12.50 11.11 0.63 0.06 TA/g 校正集 135 0.71~1.13 0.88 0.10 0.11 预测集 45 0.74~0.99 0.86 0.06 0.07 VC/mg 校正集 135 20.68~24.02 22.24 0.71 0.03 预测集 45 21.01~23.49 22.35 0.60 0.02 注:表中TA和VC数据以每100 mL果汁计. 表 2 不同波长提取方法筛选特征波长结果
检测指标 波长筛选方法 波长个数 特征波长范围/nm TSS none 1 500 1 000~2 500 CARS 118 图 2a iPLS 150 1 300~1 449 iPLS-SPA 3 1 300,1 432,1 439 TA none 1 500 1 000~2 500 CARS 74 图 2b iPLS 100 1 000~1 099 iPLS-SPA 2 1 000,1 050 VC none 1 500 1 000~2 500 CARS 41 图 2c iPLS 167 1 000~1 166 iPLS-SPA 5 1 066,1 149,1 155,1 136,1 165 表 3 建模预测结果
检测指标 波长筛选方法 LS-SVR PLSR RMSEC RC RMSEP RP RMSEC RC RMSEP RP TSS none 0.30 0.93 0.34 0.84 0.35 0.90 0.34 0.84 CARS 0.19 0.97 0.26 0.91 0.24 0.96 0.48 0.63 iPLS 0.37 0.89 0.37 0.80 0.37 0.89 0.36 0.81 iPLS-SPA 0.35 0.90 0.35 0.83 0.38 0.88 0.35 0.83 TA none 0.04 0.93 0.04 0.79 0.06 0.81 0.03 0.84 CARS 0.02 0.99 0.03 0.85 0.06 0.93 0.03 0.85 iPLS 0.04 0.90 0.04 0.80 0.06 0.81 0.03 0.84 iPLS-SPA 0.07 0.86 0.03 0.76 0.07 0.74 0.03 0.87 VC none 0.26 0.93 0.34 0.82 0.47 0.74 0.41 0.73 CARS 0.19 0.96 0.25 0.91 0.33 0.88 0.29 0.87 iPLS 0.50 0.72 0.18 0.69 0.51 0.69 0.41 0.72 iPLS-SPA 0.40 0.68 0.35 0.65 0.53 0.66 0.34 0.83 -
[1] 沈兆敏.我国柑橘产销现状、发展趋势及对策建议[J].果农之友, 2017(9): 1. doi: 10.3969/j.issn.1671-7759.2017.09.001 [2] MENDOZA F, LU R F, CEN H Y. Grading of Apples Based on Firmness and Soluble Solids Content Using Vis/SWNIR Spectroscopy and Spectral Scattering Techniques[J]. Journal of Food Engineering, 2014, 125: 59-68. doi: 10.1016/j.jfoodeng.2013.10.022 [3] FAN S X, HUANG W Q, GUO Z M, et al. Prediction of Soluble Solids Content and Firmness of Pears Using Hyperspectral Reflectance Imaging[J]. Food Analytical Methods, 2015, 8(8): 1936-1946. doi: 10.1007/s12161-014-0079-1 [4] 孙通, 莫欣欣, 刘木华.果皮对脐橙可溶性固形物可见/近红外检测精度的影响[J].光谱学与光谱分析, 2018, 38(5): 1406-1411. doi: http://d.old.wanfangdata.com.cn/Periodical_gpxygpfx201805016.aspx [5] HU M H, DONG Q L, LIU B L, et al. Prediction of Mechanical Properties of Blueberry Using Hyperspectral Interactance Imaging[J]. Postharvest Biology and Technology, 2016, 115: 122-131. doi: 10.1016/j.postharvbio.2015.11.021 [6] WANG A C, HU D, XIE L J. Comparison of Detection Modes in Terms of the Necessity of Visible Region (VIS) and Influence of the Peel on Soluble Solids Content (SSC) Determination of Navel Orange Using VIS-SWNIR Spectroscopy[J]. Journal of Food Engineering, 2014, 126: 126-132. doi: 10.1016/j.jfoodeng.2013.11.011 [7] 田喜, 何绍兰, 吕强, 等.脐橙可溶性固形物含量的光谱检测技术研究[J].西南大学学报(自然科学版), 2014, 36(3): 122-129. doi: http://xbgjxt.swu.edu.cn/article/id/jsuns2014-03-122 [8] 廖秋红, 何绍兰, 谢让金, 等.基于近红外光谱的纽荷尔脐橙产地识别研究[J].中国农业科学, 2015, 48(20): 4111-4119. doi: 10.3864/j.issn.0578-1752.2015.20.012 [9] 洪涯, 洪添胜, 代芬, 等.连续投影算法在砂糖橘总酸无损检测中的应用[J].农业工程学报, 2010, 26(S2): 380-384. doi: http://www.cqvip.com/QK/90712X/2010S2/3000101886.html [10] 王丹, 鲁晓翔, 张鹏, 等.近红外无损检测甜柿果实质地和品质[J].食品工业科技, 2013, 34(24): 53-56. doi: http://d.wanfangdata.com.cn/Periodical/spgykj201324014 [11] 王世芳, 韩平, 崔广禄, 王冬, 刘珊珊, 赵跃. 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. doi: http://www.cqvip.com/QK/90993X/20193/7001379689.html [12] 王铭海, 郭文川, 商亮, 等.基于近红外漫反射光谱的多品种桃可溶性固形物的无损检测[J].西北农林科技大学学报(自然科学版), 2014, 42(2): 142-148. doi: http://www.cqvip.com/QK/90760A/20142/49017283.html [13] GUO W C, GU J S, LIU D Y, et al. Peach Variety Identification Using Near-infrared Diffuse Reflectance Spectroscopy[J]. Computers and Electronics in Agriculture, 2016, 123: 297-303. doi: 10.1016/j.compag.2016.03.005 [14] CUNHA JUNIOR L C, NARDINI V, KHATIWADA B P, et al. Classification of Intact Açaí (Euterpe Oleracea Mart.) and Juçara (Euterpe Edulis Mart) Fruits Based on Dry Matter Content by Means of near Infrared Spectroscopy[J]. Food Control, 2015, 50: 630-636. doi: 10.1016/j.foodcont.2014.09.046 [15] 代芬, 洪添胜, 罗霞, 等.基于可见-近红外光谱的砂糖橘总酸无损检测[J].华中农业大学学报, 2012, 31(4): 518-523. doi: 10.3969/j.issn.1000-2421.2012.04.022 [16] GUO Y, NI Y N, KOKOT S. Evaluation of Chemical Components and Properties of the Jujube Fruit Using near Infrared Spectroscopy and Chemometrics[J]. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2016, 153: 79-86. doi: 10.1016/j.saa.2015.08.006 [17] LOUW E D, THERON K I. Robust Prediction Models for Quality Parameters in Japanese Plums (Prunus salicina L.) Using NIR Spectroscopy[J]. Postharvest Biology and Technology, 2010, 58(3): 176-184. doi: 10.1016/j.postharvbio.2010.07.001 [18] 夏俊芳, 李培武, 李小昱, 等.不同预处理对近红外光谱检测脐橙VC含量的影响[J].农业机械学报, 2007, 38(6): 107-111. doi: 10.3969/j.issn.1000-1298.2007.06.027 [19] 高升, 王巧华, 李庆旭, 等.基于近红外光谱的红提维生素C含量、糖度及总酸含量无损检测方法[J].分析化学, 2019, 47(6): 941-949. doi: http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=fxhx201906019 [20] 陈辰, 鲁晓翔, 张鹏, 等.红提葡萄VC含量的可见/近红外检测模型[J].食品与机械, 2015, 31(5): 70-74. doi: http://d.wanfangdata.com.cn/periodical/spyjj201505019 [21] 李雪, 刘剑飞, 张玲, 等.伏淡季水果果实内在品质综合评价[J].西南大学学报(自然科学版), 2019, 41(12): 24-32. doi: http://xbgjxt.swu.edu.cn/article/doi/10.13718/j.cnki.xdzk.2019.12.004 [22] 游庭活, 温露, 刘凡, 等.不同干燥方式及温度条件对桑椹干营养保健品质的影响[J].蚕业科学, 2015, 41(5): 921-927. [23] 詹白勺, 倪君辉, 李军.高光谱技术结合CARS算法的库尔勒香梨可溶性固形物定量测定[J].光谱学与光谱分析, 2014, 34(10): 2752-2757. doi: 10.3964/j.issn.1000-0593(2014)10-2752-06 [24] 牛晓颖, 周玉宏, 邵利敏.基于LS-SVM的草莓固酸比和可滴定酸近红外光谱定量模型[J].农业工程学报, 2013, 29(25): 270-274. doi: http://d.wanfangdata.com.cn/Periodical/nygcxb2013z1044 [25] 黄康, 汪辉君, 徐惠荣, 等.基于最小二乘支持向量机的番茄汁糖酸度分析研究[J].光谱学与光谱分析, 2009, 29(4): 931-934. doi: 10.3964/j.issn.1000-0593(2009)04-0931-04