实验室光谱仪与高光谱相机之间模型转移的研究

引用本文: 李雪莹, 任国兴, 吕美蓉, 刘岩, 孙中梁, 侯广利, 范萍萍. 实验室光谱仪与高光谱相机之间模型转移的研究[J]. 分析化学, 2021, 49(9): 1580-1586. doi: 10.19756/j.issn.0253-3820.191165 shu

Citation: LI Xue-Ying, REN Guo-Xing, LYU Mei-Rong, LIU Yan, SUN Zhong-Liang, HOU Guang-Li, FAN Ping-Ping. Study on Calibration Transfer between Laboratory Spectrometer and Hyperspectral Camera[J]. Chinese Journal of Analytical Chemistry, 2021, 49(9): 1580-1586. doi: 10.19756/j.issn.0253-3820.191165 shu

实验室光谱仪与高光谱相机之间模型转移的研究

李雪莹 ^1,2, ,
任国兴 ^1, ,
吕美蓉 ^1, ,
刘岩 ^1, ,
孙中梁 ^1, ,
侯广利 ^1, ,
范萍萍 ^{1,
,}

1.
齐鲁工业大学(山东省科学院), 山东省科学院海洋仪器仪表研究所, 青岛 266061;
2.
中国石油大学(华东) 地球科学与技术学院, 青岛 266580

通讯作者: 范萍萍,E-mail:fanpp_sdioi@126.com

基金项目:
国家自然科学基金项目（Nos.U2006209，3130041）和山东省自然科学基金项目（Nos.ZR2018LD007，ZR2017BB037）资助。

摘要: 实验室光谱仪是指通过光谱仪仅能获取光谱信号的光谱仪，高光谱相机可通过图像获取图像和光谱两种信息。在使用实验室光谱仪的条件下，针对不同温度、不同实验室的光谱仪、不同测量条件、不同地区间样品等因素的模型转移均已有研究报道，但是，针对高光谱相机与实验室光谱仪之间的模型转移问题的研究较少。本研究以青岛164份土壤样品为例，获取其实验室光谱仪光谱数据和高光谱相机数据，以实验室光谱仪光谱数据分别建立总氮（TN）、总磷（TP）含量模型，将分段直接矫正算法、模型更新法和斜率/截距修正法3种方法联用（Piecewise direct standardization-model updating-slope/bias correction method，PDS-MP-S/B），对高光谱数据进行模型转移，将转移后的高光谱数据代入TN、TP含量模型预测其含量值，评价预测效果；同时分析不同PDS窗口个数和标准集个数对预测结果的影响。在TN含量预测中，当PDS窗口个数为19、标准集个数为120时，模型转移后预测效果最好，检验集绝对系数（Absolute coefficient of test set，R_t²）为0.736，预测均方根误差（Root mean square error of prediction，RMSEP）为0.274；在TP含量预测中，当PDS窗口个数为23、标准集个数为80时，模型转移后预测效果最好，R_t²为0.647，RMSEP值为0.231。实验室光谱仪和高光谱相机之间模型转移问题的研究和解决，对高光谱相机采集的大量图像信息数据可进行快速预测，大大减少了工作量，为高光谱相机广泛应用于定量分析、快速检测技术提供了重要参考。

关键词:

English

Study on Calibration Transfer between Laboratory Spectrometer and Hyperspectral Camera

1.
Institute of Oceanographic Instrumentation, Qilu University of Technology (Shandong Academy of Sciences), Qingdao 266061, China;
2.
School of Geosciences, University of Petroleum (East China), Qingdao 266580, China

Corresponding author: FAN Ping-Ping, fanpp_sdioi@126.com

Received Date: 29 March 2019
Revised Date: 19 April 2021
Fund Project:
Supported by the National Natural Science Foundation of China (Nos.U2006209, 3130041) and the Shandong Provincial Natural Science Foundation, China (Nos.ZR2018LD007, ZR2017BB037).

Abstract: Laboratory spectrometer is a spectrometer that can only obtain spectral signals by spectrometer, and hyperspectral camera acquires image and spectral information by image. Under the condition of using laboratory spectrometer, the calibration transfer of different temperatures, different ordinary spectrometers, different measuring conditions and samples from different regions has been studied. But there are few studies on the calibration transfer between laboratory spectrometer and hyperspectral camera. Taking 164 soil samples from Qingdao as an example, the spectral data of laboratory spectrometer and hyperspectral camera data were obtained in this work. The content models of total nitrogen (TN) and total phosphorus (TP) were established based on the spectral data of laboratory spectrometer. Three methods, namely piecewise direct correction algorithm, model updating and slope/bias correction (PDS-MP-S/B), were used to transfer the hyperspectral data. The transferred hyperspectral data were substituted into the content models of TN and TP to predict their content values and evaluate the prediction effect. Meanwhile, the influence of the number of PDS window and the number of standard set on the prediction results was analyzed. When the number of PDS window was 19 and the number of standard set was 120, the prediction effect after calibration transfer was the best, absolute coefficient of test set (R_t²) was 0.736 and root mean square error of prediction (RMSEP) was 0.274. In the content prediction of TP, when the number of PDS window was 23 and the number of standard set was 80, the prediction effect after calibration transfer was the best, R_t² was 0.647 and RMSEP was 0.231. The solution of model transfer between laboratory spectrometers and hyperspectral cameras provided a powerful basis for rapid prediction of a large number of image information data collected by hyperspectral cameras, which greatly reduced the workload, and the wide application of hyperspectral cameras in quantitative analysis and rapid measurement technology.

Key words:

1. [1]
  GUO Y, NI Y, KOKOT S. Spectrochim. Acta, Part A, 2016, 153:79-86.GUO Y, NI Y, KOKOT S. Spectrochim. Acta, Part A, 2016, 153:79-86.
2. [2]
  NASEER N, QURESHI N K, NOORI F M, HONG K S. Front. Human Neurosci., 2016, 10:237.NASEER N, QURESHI N K, NOORI F M, HONG K S. Front. Human Neurosci., 2016, 10:237.
3. [3]
  LI X Y, LIU Y, LV M R, ZOU Y, FAN P P. J. Spectrosc., 2018, 2018:8513215.LI X Y, LIU Y, LV M R, ZOU Y, FAN P P. J. Spectrosc., 2018, 2018:8513215.
4. [4]
  SANGHAVI S V, MARTONCHIK J V, DAVIS A B, DINER D J. J. Quant. Spectrosc. Radiat. Transfer, 2013, 116:1-16.SANGHAVI S V, MARTONCHIK J V, DAVIS A B, DINER D J. J. Quant. Spectrosc. Radiat. Transfer, 2013, 116:1-16.
5. [5]
  BIN J, LI X, FAN W, ZHOU J H, WANG C W. Analyst, 2017, 142(12):2229-2238.BIN J, LI X, FAN W, ZHOU J H, WANG C W. Analyst, 2017, 142(12):2229-2238.
6. [6]
  ESKILDSEN CE, HANSEN PW, SKOV T, MARINI F, NØRGAARD L. J. Near Infrared Spec., 2016, 24:151-156.ESKILDSEN CE, HANSEN PW, SKOV T, MARINI F, NØRGAARD L. J. Near Infrared Spec., 2016, 24:151-156.
7. [7]
  ROOSJEN P P J, BREDE B, SUOMALAINEN J M, BARTHOLOMEUS H M, KOOISTRA L, CLEVERS J G P W. Int. J. Appl. Earth Obs., 2018, 66:14-26.ROOSJEN P P J, BREDE B, SUOMALAINEN J M, BARTHOLOMEUS H M, KOOISTRA L, CLEVERS J G P W. Int. J. Appl. Earth Obs., 2018, 66:14-26.
8. [8]
  PU Y Y, SUN D W, RICCIOLI C, BUCCHERI M, GRASSI M, CATTANEO T M P, GOWEN A. Food Anal. Methods, 2018, 11(4):1021-1033.PU Y Y, SUN D W, RICCIOLI C, BUCCHERI M, GRASSI M, CATTANEO T M P, GOWEN A. Food Anal. Methods, 2018, 11(4):1021-1033.
9. [9]
  FERNANDEZ L, GUNEY S, GUTIERREZ-GALVEZ A, MARCO S. Sens. Actuators, B, 2016, 231:276-284.FERNANDEZ L, GUNEY S, GUTIERREZ-GALVEZ A, MARCO S. Sens. Actuators, B, 2016, 231:276-284.
10. [10]
  MILANEZ K D T M, NÓBREGA T C A, NASCIMENTO D S, INSAUSTI M, PONTES M J C. Microchem. J., 2017, 133:669-675.MILANEZ K D T M, NÓBREGA T C A, NASCIMENTO D S, INSAUSTI M, PONTES M J C. Microchem. J., 2017, 133:669-675.
11. [11]
  ZHANG J, ZHANG Z, XIANG Y, DAI Y, HARRINGTON P D B. Talanta, 2011, 83(5):1401-1409.ZHANG J, ZHANG Z, XIANG Y, DAI Y, HARRINGTON P D B. Talanta, 2011, 83(5):1401-1409.
12. [12]
  PEREIRA L S, CARNEIRO M F, BOTELHO B G, SENA M M. Talanta, 2016, 147:351-357.PEREIRA L S, CARNEIRO M F, BOTELHO B G, SENA M M. Talanta, 2016, 147:351-357.
13. [13]
  NOURI M, GOMEZ C, GORRETTA N, ROGER J M. Geoderma, 2017, 298:54-66.NOURI M, GOMEZ C, GORRETTA N, ROGER J M. Geoderma, 2017, 298:54-66.
14. [14]
  COOPER J B, LARKIN C M, ABDELKADER M F. J. Near Infrared Spectrosc., 2011, 19(2):139-150.COOPER J B, LARKIN C M, ABDELKADER M F. J. Near Infrared Spectrosc., 2011, 19(2):139-150.

扫一扫看文章

计量

PDF下载量: 17
文章访问数: 1146
HTML全文浏览量: 319

通讯作者: 陈斌, bchen63@163.com

1.
沈阳化工大学材料科学与工程学院沈阳 110142

实验室光谱仪与高光谱相机之间模型转移的研究

通讯作者: 范萍萍,E-mail:fanpp_sdioi@126.com

English

Study on Calibration Transfer between Laboratory Spectrometer and Hyperspectral Camera

Corresponding author: FAN Ping-Ping, fanpp_sdioi@126.com

CCS Chemistry：嵌段共聚物自组装成 “三明治”二维有机材料，实现纳米级压力传感功能

CCS Chemistry：颜徐州、鲍哲南：你的皮肤可能是一片SPMs

CCS Chemistry：工作高效不疲劳，只须让催化剂动起来

CCS Chemistry：静电组装导向聚合- 聚电解质纳米凝胶量化制备新策略

CCS Chemistry：金黄色葡萄球菌醛基脱氢酶是如何识别特异性底物的？

计量

通讯作者: 陈斌, bchen63@163.com

中国化学会秘书处

实验室光谱仪与高光谱相机之间模型转移的研究

通讯作者: 范萍萍,E-mail:fanpp_sdioi@126.com

English

Study on Calibration Transfer between Laboratory Spectrometer and Hyperspectral Camera

Corresponding author: FAN Ping-Ping, fanpp_sdioi@126.com

CCS Chemistry：嵌段共聚物自组装成 “三明治”二维有机材料，实现纳米级压力传感功能

CCS Chemistry：颜徐州、鲍哲南： 你的皮肤可能是一片SPMs

CCS Chemistry：工作高效不疲劳，只须让催化剂动起来

CCS Chemistry：静电组装导向聚合- 聚电解质纳米凝胶量化制备新策略

CCS Chemistry：金黄色葡萄球菌醛基脱氢酶是如何识别特异性底物的？

计量

文章相关

通讯作者: 陈斌, bchen63@163.com

中国化学会秘书处

CCS Chemistry：颜徐州、鲍哲南：你的皮肤可能是一片SPMs