留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于高光谱成像技术的涌泉蜜桔糖度最优检测位置

李斌 万霞 刘爱伦 邹吉平 卢英俊 姚迟 刘燕德

李斌, 万霞, 刘爱伦, 邹吉平, 卢英俊, 姚迟, 刘燕德. 基于高光谱成像技术的涌泉蜜桔糖度最优检测位置[J]. 中国光学(中英文), 2024, 17(1): 128-139. doi: 10.37188/CO.2023-0057
引用本文: 李斌, 万霞, 刘爱伦, 邹吉平, 卢英俊, 姚迟, 刘燕德. 基于高光谱成像技术的涌泉蜜桔糖度最优检测位置[J]. 中国光学(中英文), 2024, 17(1): 128-139. doi: 10.37188/CO.2023-0057
LI Bin, WAN Xia, LIU Ai-lun, ZOU Ji-ping, LU Ying-jun, YAO Chi, LIU Yan-de. Optimal position for suger content detection of Yongquan honey oranges based on hyperspectral imaging technology[J]. Chinese Optics, 2024, 17(1): 128-139. doi: 10.37188/CO.2023-0057
Citation: LI Bin, WAN Xia, LIU Ai-lun, ZOU Ji-ping, LU Ying-jun, YAO Chi, LIU Yan-de. Optimal position for suger content detection of Yongquan honey oranges based on hyperspectral imaging technology[J]. Chinese Optics, 2024, 17(1): 128-139. doi: 10.37188/CO.2023-0057

基于高光谱成像技术的涌泉蜜桔糖度最优检测位置

基金项目: 青年科学基金项目(No.12103019)
详细信息
    作者简介:

    李 斌(1989-),男,江西鹰潭人,博士,副教授,华东交通大学机电学院教师,2012年于武汉大学获得学士学位,2017年于中国科学院光电技术研究所获得博士学位,主要研究方向:无损智能检测。E-mail:libingioe@126.com

    刘燕德(1967—),女,江西吉安人,博士,二级教授,智能机电装备创新研究院院长,2001 年于江西农业大学获得硕士学位,2006 年于浙江大学获得博士学位,主要研究方向:光电无损检测技术、光机电一体化技术与装备。E-mail:liuyande@ecjtu.jx.cn

  • 中图分类号: O433.4

Optimal position for suger content detection of Yongquan honey oranges based on hyperspectral imaging technology

Funds: Supported by Youth Science Fund Projects (No.12103019)
More Information
  • 摘要:

    本文旨在探索涌泉蜜桔糖度的最优检测位置和最佳预测模型,以便为蜜桔糖度检测分级提供理论依据。本文利用波长为390.2~981.3 nm的高光谱成像系统对涌泉蜜桔糖度最佳检测位置进行研究,将涌泉蜜桔的花萼、果茎、赤道和全局的光谱信息与其对应部位的糖度结合,建立其预测模型。使用标准正态变量变换(SNV)、多元散射校正(MSC)、基线校准(Baseline)和SG平滑(Savitzkv-Golay)4种预处理方法对不同部位的原始光谱进行预处理,用预处理后的光谱数据建立偏最小二乘回归(PLSR)和最小二乘支持向量机(LSSVM)模型。找出蜜桔不同部位的最佳预处理方式,对经过最佳预处理后的光谱数据采用竞争性自适应重加权算法(CARS)和无信息变量消除法(UVE)进行特征波长筛选。最后,用筛选后的光谱数据建立PLSR和LSSVM模型并进行分析比较。研究结果表明,全局的MSC-CARS-LSSVM模型预测效果最佳,其预测集相关系数Rp=0.955,均方根误差RMSEP=0.395,其次是蜜桔赤道部位的SNV-PLSR模型,其预测集相关系数Rp=0.936,均方根误差RMSEP=0.37。两者预测集相关系数相近,因此可将赤道位置作为蜜桔糖度的最优检测位置。本研究表明根据蜜桔不同部位建立的糖度预测模型的预测效果有所差异,研究最优检测位置和最佳预测模型可以为蜜桔进行糖度检测分级提供理论依据。

     

  • 图 1  涌泉蜜桔不同部位图像

    Figure 1.  Images of different parts of Yongquan honey oranges

    图 2  高光谱成像装置示意图

    Figure 2.  Schematic diagram of the hyperspectral imaging device

    图 3  涌泉蜜桔光谱曲线。(a)不同部位的原始光谱曲线;(b)不同部位的平均光谱曲线

    Figure 3.  Spectral curves of Yongquan honey orange. (a) Original spectral curves of different parts; (b) average spectral curves of different parts

    图 4  CARS果茎部位特征波长选择过程。(a)变量数变化;(b)交叉验证均方根变化;(c)回归系数变化

    Figure 4.  Selecting process of the characteristic wavelength of the fruit stem part by CARS. (a) Changes in number of variables; (b) changes in the RMSECV; (c) changes in regression coefficient

    图 5  两种预处理方法基于CARS算法果茎部位特征波长位置图。(a)Baseline;(b)MSC

    Figure 5.  Location map of the characteristic wavelengths in the fruit stem part based on the CARS algorithm corresponding to the pretreatments (a) Baseline and (b) MSC

    图 6  各部位基于CARS算法特征波长位置图。(a)花萼;(b)赤道;(c)全局

    Figure 6.  Location maps of the characteristic wavelengths based on CARS algorithm. (a) Calyx; (b) equator and (c) global

    图 7  UVE筛选后果茎部位的稳定性值图

    Figure 7.  Stability values of the fruit stem part after UVE screening

    图 8  两种预处理方法下,基于UVE算法果茎部位特征波长位置图。(a)Baseline;(b)MSC

    Figure 8.  Location maps of characteristic wavelengths in the fruit stem part based on the UVE algorithm corresponding to (a) Baseline and (b) MSC

    图 9  基于UVE算法特征波长位置图。(a)花萼;(b)赤道;(c)全局

    Figure 9.  Location maps of the characteristic wavelengths of the UVE-based algorithm. (a) Calyx; (b) equator and (c) global

    图 10  涌泉蜜桔糖度含量预测模型(a)MSC-CARS-PLSR和(b)MSC-CARS-LSSVM的散点图

    Figure 10.  Scatter plots of the Yongquan honey oranges sugar content prediction models (a) MSC-CARS-PLSR and (b) MSC-CARS-LSSVM

    表  1  涌泉蜜桔不同部位的糖度统计分析结果

    Table  1.   Statistical analysis of the sugar content of different parts of Yongquan honey orange

    蜜桔部位样本数最大值
    / OBrix
    最小值
    / OBrix
    平均值
    / OBrix
    标准差
    / OBrix
    花萼12019.810.815.21.39
    果茎12017.910.114.21.52
    赤道12018.211.314.51.37
    全局12017.811.214.61.34
    下载: 导出CSV

    表  2  基于不同预处理方法的涌泉蜜桔糖度检测PLSR模型比较

    Table  2.   Comparison of PLSR models for detecting the sugar content of Yongquan honey orange based on different pretreatments

    预测模型预处理方法建模集预测集
    RCRMSEC/OBrixRPRMSEP/OBrix
    花萼模型Raw0.9460.3840.8930.457
    SNV0.8470.580.8060.688
    MSC0.8320.6220.7660.564
    Baseline0.9210.4090.8900.518
    SG0.9320.4270.8980.436
    果茎模型Raw0.9490.4280.8590.587
    SNV0.9020.5930.8820.669
    MSC0.8890.5990.8640.587
    Baseline0.9310.4980.9130.468
    SG0.9430.4550.8680.569
    赤道模型Raw0.9320.4710.8610.553
    SNV0.9460.4080.9360.370
    MSC0.9600.3650.8780.458
    Baseline0.9640.3490.9330.384
    SG0.9240.4970.8610.555
    全局模型Raw0.9710.3050.9200.388
    SNV0.9450.4030.9010.435
    MSC0.9530.3740.9340.435
    Baseline0.9260.4690.8550.495
    SG0.9270.4760.9230.384
    下载: 导出CSV

    表  3  基于不同预处理的涌泉蜜桔糖度LSSVM模型比较

    Table  3.   Comparison of LSSVM models for detecting the sugar content of Yongquan honey orange basedon different pretreatments

    预测模型预处理
    方法
    建模集预测集
    RCRMSEC/OBrixRPRMSEP/ OBrix
    花萼模型Raw0.9210.4700.8600.513
    SNV0.9380.3830.7890.700
    MSC0.9590.3230.7880.539
    Baseline0.9420.3600.8690.585
    SG0.9230.4590.8760.477
    果茎模型Raw0.9790.2860.7820.750
    SNV0.9080.5940.8340.710
    MSC0.9550.4040.8840.596
    Baseline0.9240.5270.6420.854
    SG0.9530.4190.8270.650
    赤道模型Raw0.9650.3550.8290.594
    SNV0.9540.3880.9060.405
    MSC0.9730.3150.8270.530
    Baseline0.9790.2810.8670.544
    SG0.9560.3880.8450.575
    全局模型Raw0.9620.3550.8920.443
    SNV0.9720.2960.8970.456
    MSC0.9800.2530.9460.400
    Baseline0.9730.2930.8110.590
    SG0.9610.3560.9090.414
    下载: 导出CSV

    表  4  基于CARS特征波长筛选后蜜桔不同部位的PLSR和LSSVM模型比较

    Table  4.   Comparison of PLSR and LSSVM models for different parts of honey oranges after CARS characteristic wavelengths screening

    预测模型不同部位建模集预测集
    RCRMSEC/ OBrixRPRMSEP/ OBrix
    PLSR花萼0.9260.4470.9180.400
    果茎0.9280.5070.9220.424
    赤道0.9330.4520.9140.400
    全局0.9480.3940.9420.399
    LSSVM花萼0.9270.4450.9140.408
    果茎0.9510.4120.9040.546
    赤道0.9600.3520.9010.423
    全局0.9750.2740.9550.395
    下载: 导出CSV

    表  5  基于UVE特征波长筛选后蜜桔不同部位的PLSR和LSSVM模型比较

    Table  5.   Comparison of PLSR and LSSVM models for different parts of honey oranges after UVE characteristic wavelengths screening

    预测模型不同部位建模集预测集
    RCRMSEC/ OBrixRPRMSEP/ OBrix
    PLSR花萼0.8900.5380.8500.519
    果茎0.8850.6330.8120.655
    赤道0.9430.4190.9330.364
    全局0.9490.3930.9370.434
    LSSVM花萼0.9010.5140.8380.537
    果茎0.9500.4160.8960.575
    赤道0.9500.4000.9000.423
    全局0.9560.3680.9430.414
    下载: 导出CSV
  • [1] 王玲娇, 蒋守渭. 临海涌泉蜜桔网络营销策略研究[J]. 现代商业,2015(35):36-37.

    WANG L J, JIANG SH W. Research on network marketing strategy of Linhai Yongquan honey tangerine[J]. Modern Business, 2015(35): 36-37. (in Chinese)
    [2] 介邓飞, 杨杰, 彭雅欣, 等. 基于高光谱技术的柑橘不同部位糖度预测模型研究[J]. 食品与机械,2017,33(3):51-54.

    JIE D F, YANG J, PENG Y X, et al. Research on the detection model of sugar content in different position of citrus based on the hyperspectral technology[J]. Food &Machinery, 2017, 33(3): 51-54. (in Chinese)
    [3] 王玥, 樊柳荫. 猪肉可视化新鲜度智能指示薄膜研究[J]. 分析化学,2023,51(1):139-145.

    WANG Y, FAN L Y. Intelligent indicator film for visual meat freshness monitoring[J]. Chinese Journal of Analytical Chemistry, 2023, 51(1): 139-145. (in Chinese)
    [4] 田喜, 陈立平, 王庆艳, 等. 全透射近红外光谱的苹果整果糖度在线检测模型优化[J]. 光谱学与光谱分析,2022,42(6):1907-1914.

    TIAN X, CHEN L P, WANG Q Y, et al. Optimization of online determination model for sugar in a whole apple using full transmittance spectrum[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1907-1914. (in Chinese)
    [5] 陈玥瑶, 夏静静, 韦芸, 等. 近红外光谱法无损检测平谷产大桃品质方法研究[J]. 分析化学,2023,51(3):454-462.

    CHEN Y Y, XIA J J, WEI Y, et al. Research on nondestructive quality test of Pinggu peach by near-infrared spectroscopy[J]. Chinese Journal of Analytical Chemistry, 2023, 51(3): 454-462. (in Chinese)
    [6] LU Y ZH, HUANG Y P, LU R F. Innovative hyperspectral imaging-based techniques for quality evaluation of fruits and vegetables: A review[J]. Applied Sciences, 2017, 7(2): 189. doi: 10.3390/app7020189
    [7] MESA A R, CHIANG J Y. Multi-input deep learning model with RGB and hyperspectral imaging for banana grading[J]. Agriculture, 2021, 11(8): 687. doi: 10.3390/agriculture11080687
    [8] 徐婧, 郑红, 谢丽芳, 等. 鸡肉样品中痕量喹诺酮类抗生素的表面增强拉曼光谱检测研究[J]. 分析化学,2023,51(3):397-404.

    XU J, ZHENG H, XIE L F, et al. Fast detection of trace enrofloxacin and ciprofloxacin in chicken meat by surface-enhanced Raman spectroscopy[J]. Chinese Journal of Analytical Chemistry, 2023, 51(3): 397-404. (in Chinese)
    [9] 沈彦龙, 程立业, 孟祥茹, 等. 人参连作土壤对不同生育期人参生长发育及抗氧化系统的影响[J]. 应用化学,2023,40(1):109-115.

    SHEN Y L, CHENG L Y, MENG X R, et al. Effects of ginseng continuous soil crop on growth development and antioxidant system of ginseng at different fertility stages[J]. Chinese Journal of Applied Chemistry, 2023, 40(1): 109-115. (in Chinese)
    [10] 黄蕊, 叶长青, 李亚军, 等. 线粒体靶向的近红外HClO/ClO-荧光探针的研究进展[J]. 应用化学,2022,39(3):407-424.

    HUANG R, YE CH Q, LI Y J, et al. Progress of mitochondria-targeted near-infrared HClO/ClO- fluorescent probes[J]. Chinese Journal of Applied Chemistry, 2022, 39(3): 407-424. (in Chinese)
    [11] 王瑞, 孟祥茹, 李琼, 等. 人参属中药腐解化感作用研究进展[J]. 应用化学,2023,40(1):1-8.

    WANG R, MENG X R, LI Q, et al. Research progress on the decomposed Allelopathy of Panax genus[J]. Chinese Journal of Applied Chemistry, 2023, 40(1): 1-8. (in Chinese)
    [12] KIM D, BURKS T F, RITENOUR M A, et al. Citrus black spot detection using hyperspectral imaging[J]. International Journal of Agricultural and Biological Engineering, 2014, 7(6): 20-27.
    [13] 孟田源, 王转卫, 迟茜, 等. 基于高光谱成像技术生长发育后期苹果糖度的无损检测[J]. 西北农林科技大学学报(自然科学版),2016,44(6):228-234.

    MENG T Y, WANG ZH W, CHI X, et al. Hyperspectral imaging based non-destructive prediction of soluble solids content in apples at late development period[J]. Journal of Northwest A&F University (Natural Science Edition), 2016, 44(6): 228-234. (in Chinese)
    [14] 许丽佳, 陈铭, 王玉超, 等. 高光谱成像的猕猴桃糖度无损检测方法[J]. 光谱学与光谱分析,2021,41(7):2188-2195.

    XU L J, CHEN M, WANG Y CH, et al. Study on non-destructive detection method of kiwifruit sugar content based on hyperspectral imaging technology[J]. Spectroscopy and Spectral Analysis, 2021, 41(7): 2188-2195. (in Chinese)
    [15] 杨杰, 马本学, 王运祥, 等. 葡萄可溶性固形物的高光谱无损检测技术[J]. 江苏农业科学,2016,44(6):401-403.

    YANG J, MA B X, WANG Y X, et al. Hyperspectral nondestructive detection technique of soluble solids in grapes[J]. Jiangsu Agricultural Science, 2016, 44(6): 401-403. (in Chinese)
    [16] SUN X D, XU C, LUO CH G, et al. Non-destructive detection of tea stalk and insect foreign bodies based on THz-TDS combination of electromagnetic vibration feeder[J]. Food Quality and Safety, 2023, 7: fyad004. doi: 10.1093/fqsafe/fyad004
    [17] LIU Y D, SUN X D, OUYANG A G. Nondestructive measurement of soluble solid content of navel orange fruit by visible-NIR spectrometric technique with PLSR and PCA-BPNN[J]. LWT - Food Science and Technology, 2010, 43(4): 602-607. doi: 10.1016/j.lwt.2009.10.008
    [18] YANG Y, ZHAO CH J, HUANG W Q, et al. Optimization and compensation of models on tomato soluble solids content assessment with online Vis/NIRS diffuse transmission system[J]. Infrared Physics &Technology, 2022, 121: 104050.
    [19] 袁琳, 徐怀德, 李钰金. 近红外漫反射光谱检测网纹瓜可溶性固形物含量的研究[J]. 中国食品学报,2010,10(4):272-277.

    YUAN L, XU H D, LI Y J. Studies on the rapid measurements of soluble solids content in nutmeg melon by near infrared diffuse reflectance spectroscopy[J]. Journal of Chinese Institute of Food Science and Technology, 2010, 10(4): 272-277. (in Chinese)
    [20] ZHANG D Y, XU L, WANG Q Y, et al. The optimal local model selection for robust and fast evaluation of soluble solid content in melon with thick peel and large size by Vis-NIR spectroscopy[J]. Food Analytical Methods, 2019, 12(1): 136-147. doi: 10.1007/s12161-018-1346-3
    [21] 孙博康, 刘贵珊. 基于高光谱技术检测香水梨硬度的研究[J]. 食品安全导刊,2021(19):118-120.

    SUN B K, LIU G SH. Study on the detection of hardness of perfumed pears based on hyperspectral technology[J]. Journal of Food Safety, 2021(19): 118-120. (in Chinese)
    [22] ZHANG F, LI B, YIN H, et al. Study on the quantitative assessment of impact damage of yellow peaches using the combined hyperspectral technology and mechanical parameters[J]. Journal of Spectroscopy, 2022, 2022: 7526826.
    [23] WANG ZH L, CHEN J, FAN Y F, et al. Evaluating photosynthetic pigment contents of maize using UVE-PLS based on continuous wavelet transform[J]. Computers and Electronics in Agriculture, 2020, 169: 105160. doi: 10.1016/j.compag.2019.105160
    [24] 廉小亲, 汤燊淼, 吴静珠, 等. 基于近红外的兰州百合品质定量建模方法研究[J]. 食品科技,2020,45(7):298-302.

    LIAN X Q, TANG S M, WU J ZH, et al. Research on quantitative model of Lilium Lanzhou quality by near infrared spectroscopy[J]. Food Science and Technology, 2020, 45(7): 298-302. (in Chinese)
    [25] YANG X Y, LIU G SH, HE J G, et al. Determination of sugar content in Lingwu jujube by NIR–hyperspectral imaging[J]. Journal of Food Science, 2021, 86(4): 1201-1214. doi: 10.1111/1750-3841.15674
    [26] 李斌, 邹吉平, 张烽, 等. 基于高光谱成像技术和力学参数对贡梨冲击损伤的定量研究[J]. 中国农业大学学报,2023,28(2):186-197.

    LI B, ZOU J P, ZHANG F, et al. Quantitative study on impact damage of Gongli based on hyperspectral imaging technology and mechanical parameters[J]. Journal of China Agricultural University, 2023, 28(2): 186-197. (in Chinese)
  • 加载中
图(10) / 表(5)
计量
  • 文章访问数:  244
  • HTML全文浏览量:  153
  • PDF下载量:  199
  • 被引次数: 0
出版历程
  • 收稿日期:  2023-03-30
  • 修回日期:  2023-04-19
  • 网络出版日期:  2023-07-13

目录

    /

    返回文章
    返回