Brix, pH and anthocyanin content determination in whole Port wine grape berries by hyperspectral imaging and neural networks

被引:61
作者
Fernandes, Armando M. [1 ]
Franco, Camilo [2 ]
Mendes-Ferreira, Ana [3 ]
Mendes-Faia, Arlete [3 ]
da Costa, Pedro Leal [4 ]
Melo-Pinto, Pedro [5 ,6 ]
机构
[1] Univ Lisbon, Inst Super Tecn, IDMEC LAETA, Ctr Intelligent Syst, P-1049001 Lisbon, Portugal
[2] Univ Copenhagen, Inst Food & Resource Econ, DK-1168 Copenhagen, Denmark
[3] Univ Tras Os Montes & Alto Douro, Ctr Genom & Biotechnol, Inst Biotechnol & Bioengn, P-5001801 Vila Real, Portugal
[4] Symington Family Estates, P-4431901 Vila Nova De Gaia, Portugal
[5] Univ Tras Os Montes & Alto Douro, CITAB Ctr Res & Technol Agroenvironm & Biol Sci, P-5000801 Vila Real, Portugal
[6] Univ Tras Os Montes & Alto Douro, Escola Ciencias & Tecnol, Dept Engn, P-5000801 Vila Real, Portugal
关键词
Grapes; Touriga Franca; Spectroscopy; Neural networks; Hyperspectral; ACTIVE ODORANTS; MATURITY; SPECTROSCOPY; BRANCELLAO; SHOULDERS; CLUSTER; TIPS;
D O I
10.1016/j.compag.2015.05.013
中图分类号
S [农业科学];
学科分类号
09 ;
摘要
This work presents the results of measuring pH, sugars, and anthocyanin content of whole grape berries. The spectrum of each sample, composed of six whole grape berries, was collected using hyperspectral imaging in reflectance mode from 380 to 1028 nm. The spectra were converted to enological parameters by multilayer perceptrons created using 240 samples that were split for 7-fold cross-validation and test. The test set with 30 samples revealed R-2 values of 0.73, 0.92 and 0.95 and RMSE of 0.18, 0.95 degrees Brix and 14 mg/l for pH, sugars and anthocyanin content, respectively. This is the only work, up to the authors' knowledge, that reports the simultaneous determination of pH, sugars and anthocyanin with a spatial resolution smaller than the bunch size, by spectroscopy measurements done in reflectance mode. The results presented show an improvement of the state-of-the-art, for both reflectance and transmittance mode, that is particularly notorious for anthocyanin measurement. (C) 2015 Elsevier B.V. All rights reserved.
引用
收藏
页码:88 / 96
页数:9
相关论文
共 27 条
[1]   Maturity, variety and origin determination in white grapes (Vitis Villifera L.) using near infrared reflectance technology [J].
Arana, I ;
Jarén, C ;
Arazuri, S .
JOURNAL OF NEAR INFRARED SPECTROSCOPY, 2005, 13 (06) :349-357
[2]  
Bishop CM, 1995, Neural Networks for Pattern Recognition
[3]   Soluble solids content and pH prediction and varieties discrimination of grapes based on visible-near infrared spectroscopy [J].
Cao, Fang ;
Wu, Di ;
He, Yong .
COMPUTERS AND ELECTRONICS IN AGRICULTURE, 2010, 71 :S15-S18
[4]  
Carbonneau A., 1993, Nouveaux systemes de culture integre du vignoble
[5]   Predicting the anthocyanin content of wine grapes by NIR hyperspectral imaging [J].
Chen, Shanshan ;
Zhang, Fangfang ;
Ning, Jifeng ;
Liu, Xu ;
Zhang, Zhenwen ;
Yang, Shuqin .
FOOD CHEMISTRY, 2015, 172 :788-793
[6]   Prediction of colour and pH in grapes using a diode array spectrophotometer (400-1100 nm) [J].
Cozzolino, D ;
Esler, MB ;
Dambergs, RG ;
Cynkar, WU ;
Boehm, DR ;
Francis, IL ;
Gishen, M .
JOURNAL OF NEAR INFRARED SPECTROSCOPY, 2004, 12 (02) :105-111
[7]   Analysis of grapes and wine by near infrared spectroscopy [J].
Cozzolino, D. ;
Dambergs, R. G. ;
Janik, L. ;
Cynkar, W. U. ;
Gishen, M. .
JOURNAL OF NEAR INFRARED SPECTROSCOPY, 2006, 14 (05) :279-289
[8]  
Cozzolino D., 2005, 12 AUSTR WIN IND TEC, P334
[9]  
Fadock M., 2011, Non-Destructive VIS-NIR Reflectance Spectrometry For Red Wine Grapes Analysis
[10]   Determination of anthocyanin concentration in whole grape skins using hyperspectral imaging and adaptive boosting neural networks [J].
Fernandes, Armando Manuel ;
Oliveira, Paula ;
Moura, Joao Paulo ;
Oliveira, Ana Alexandra ;
Falco, Virgilio ;
Correia, Maria Jose ;
Melo-Pinto, Pedro .
JOURNAL OF FOOD ENGINEERING, 2011, 105 (02) :216-226