Time filter

Source Type

Tongzhou District, China

Cheng J.-H.,South China University of Technology | Sun D.-W.,South China University of Technology | Sun D.-W.,University College Dublin | Qu J.-H.,South China University of Technology | And 5 more authors.
Journal of Food Engineering

This study investigated the feasibility of developing a multispectral imaging method using key wavelengths from hyperspectral images for modeling and simultaneously predicting total volatile basic nitrogen (TVB-N), thiobarbituric acid reactive substances (TBARS) and K value in grass carp fillet during chemical spoilage. The established least-squares support vector machine (LS-SVM) and multiple linear regression (MLR) models using five successive projection algorithm (SPA)-selected and six genetic algorithm (GA)-selected wavelengths showed excellent performances for predicting TVB-N and K value with R2 P > 0.900 and RPD > 3.000, and poor results for TBARS value prediction. The LS-SVM model using six GA-selected wavelengths showed good reliability and was considered the best for simultaneous determination of TVB-N, TBARS and K value. The distribution maps of chemical spoilage changes were generated using image processing algorithms. The results demonstrated the feasibility of developing a rapid and on-line multispectral imaging system using the feature wavelengths and chemometrics analysis. © 2016 Elsevier Ltd. Source

Sun M.,Beijing Technology and Business University | Chen X.,Zolix Instruments Co. | Zhang H.,Zolix Instruments Co. | Chen H.,Zolix Instruments Co.
Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering

Image cubes containing continuous spectral waveband information, in which the image information could be used for external attribute inspection while the spectral information could be applied to the internal attribute inspection, could be obtained from implementing a hyperspectral image technology which combines the advantages of computer vision and spectroscopy. Apples were adopted as the experimental object. A hyperspectral imaging system with the wavelength range of 400-1000 nm was built for detecting bruises. The hyperspectral imaging system was used as a powerful tool to determine the effective wavelengths that could be used for the detection of bruises on apples. The optimal wavelength region 550-950 nm for bruise detection was selected by the Principal component analysis (PCA), which is a very effective method for data dimension reduction and feature extraction of the hyperspectral data cube. The effective wavelengths 714 nm with weighing coefficients at peaks was determined using the loading coefficients of the PC4 image of PCA on 400-1000 nm. Source

Cheng J.-H.,South China University of Technology | Sun D.-W.,South China University of Technology | Sun D.-W.,University College Dublin | Pu H.-B.,South China University of Technology | And 4 more authors.
Journal of Food Engineering

The investigation of visible and near infrared hyperspectral imaging (400-1000 nm) coupled with classifiers and spectral pre-processing techniques was conducted to discriminate fresh from cold-stored (4 °C for 7 days) and frozen-thawed (-20 °C and -40 °C for 30 days) grass carp fish fillets. Four classifiers with four spectral pretreatment methods were applied to establish the classification models. Compared with the original models established using the full wavelengths, the classification models with three classifiers of soft independent modeling of class analogy (SIMCA), least squares-support vector machine (LS-SVM) and probabilistic neural network (PNN) in tandem with the first derivative pretreatment showed the best classification performance and the highest correct classification rate (CCR) of 94.29%. In addition, in order to reduce the high dimensionality of hyperspectral images, seven optimal wavelengths were selected by successive projections algorithm (SPA) and used to simplify the classification models. The simplified model obtained by the LS-SVM classifier coupled with the first derivative pre-processing method also presented good prediction accuracy with the CCR of 91.43%. The results demonstrated that the integration of hyperspectral imaging and classifiers analysis had a great potential for on-line detection and was feasible to rapidly and non-invasively discriminate fresh and frozen-thawed fish fillets. © 2015 Elsevier Ltd. All rights reserved. Source

Discover hidden collaborations