Search

Spectral data from multiple sources can be integrated into multi-block fusion chemometric models, such as sequentially orthogonalized partial-least squares (SO-PLS), to improve the prediction of sample quality features. Pre-processing techniques are often applied to mitigate extraneous variability, unrelated to the response variables. However, the selection of suitable pre-processing methods and identification of informative data blocks becomes increasingly complex and time-consuming when dealing with a large number of blocks. The problem addressed in this work is the efficient pre-processing, selection, and ordering of data blocks for targeted applications in SO-PLS. We introduce the PROSAC–SO–PLS methodology, which employs pre-processing ensembles with response-oriented sequential alternation calibration (PROSAC). This approach identifies the best pre-processed data blocks and their sequential order for specific SO-PLS applications. The method uses a stepwise forward selection strategy, facilitated by the rapid Gram-Schmidt process, to prioritize blocks based on their effectiveness in minimizing prediction error, as indicated by the lowest prediction residuals. To validate the efficacy of our approach, we showcase the outcomes of three empirical near-infrared (NIR) datasets. Comparative analyses were performed against partial-least-squares (PLS) regressions on single-block pre-processed datasets and a methodology relying solely on PROSAC. The PROSAC–SO–PLS approach consistently outperformed these methods, yielding significantly lower prediction errors. This has been evidenced by a reduction in the root-mean-squared error of prediction (RMSEP) ranging from 5 to 25 % across seven out of the eight response variables analyzed. The PROSAC–SO–PLS methodology offers a versatile and efficient technique for ensemble pre-processing in NIR data modeling. It enables the use of SO-PLS minimizing concerns about pre-processing sequence or block order and effectively manages a large number of data blocks. This innovation significantly streamlines the data pre-processing and model-building processes, enhancing the accuracy and efficiency of chemometric models. Summary of the PROSAC–SO–PLS methodology applied to one of the response variables (protein) of the miniS-milk dataset, which is one of the three datasets presented in this study. In this specific case, multiple blocks of spectral data derived from multiple miniature spectrometers measuring the same milk samples simultaneously are subjected to an ensemble of pre-processing techniques within the PROSAC framework. The five most variance-explaining blocks identified by PROSAC are subsequently utilized as input variables for SO-PLS, facilitating the construction of a model that optimizes block utilization for predicting the protein content in milk, with protein content reference values obtained from laboratory analyses. [Display omitted] • PROSAC and SO-PLS are combined for improved multi-block data fusion. • PROSAC streamlines SO-PLS by easing pre-processing and block order concerns. • PROSAC–SO–PLS effectively manages a large number of input data blocks. • PROSAC–SO–PLS is tested on three NIR datasets, surpassing existing state-of-art. • PROSAC–SO–PLS reduces prediction errors by 5–25 % for the evaluated NIR datasets. [ABSTRACT FROM AUTHOR]

• Reflectance, Transmittance, Fluorescence, Raman and integrated imaging is covered. • Instrumentation is discussed with special attention for real-time imaging. • The major steps in hyperspectral data analysis are discussed. • Overview of applications in external and internal quality evaluation. • Review of applications in food safety detection. In the past 20 years, hyperspectral imaging has been widely investigated as an emerging, promising technology for evaluating quality and safety of horticultural products. This technology has originated from remote sensing and joins the domains of machine vision and point spectroscopy to provide superior image segmentation for the detection of defects and contaminations, and to map the chemical composition. Thanks to the advancements in instrumentation and data analysis in the past two decades, hyperspectral imaging technology has evolved into a powerful nondestructive inspection tool and the scope of applications in postharvest quality and safety evaluation has expanded tremendously. In this article, different imaging modes (reflectance, transmittance, fluorescence and Raman) and their combinations, and the potential for real-time acquisition of hyperspectral images at industry relevant speeds are first discussed in terms of their advantages and disadvantages. Next reviewed are different data processing/analysis methods and associated steps from data pre-processing over the spectral and spatial domains to the actual model building and performance evaluation. An overview is then given of hyperspectral imaging applications for external quality and defect evaluation, internal quality and maturity assessment, and food safety detection of horticultural products. Finally, a brief discussion is presented on the challenges and opportunities in future development and application of hyperspectral imaging technology in food quality and safety evaluation of horticultural products. [ABSTRACT FROM AUTHOR]

Temperature fluctuation commonly occurs in the cold chain leading to complete or partial thawing and refreezing of frozen products resulting in a multifrozen product. Such oscillation of temperature could cause significant quality reduction compared to single frozen products. This study was designed to differentiate frozen Atlantic salmon fillets based on the level of temperature fluctuation. Near‐infrared spectroscopy (NIRS) coupled with chemometrics was used to classify the frozen fillets stored at no fluctuation (NF), low fluctuation (LF), high fluctuation (HF), and very high fluctuation (VF) temperature. Using spectral profiles obtained at both frozen and thawed states, fillets were classified based on the level of temperature fluctuation by partial least squares discriminant analysis (PLS‐DA). The thawed samples showed better classification accuracy (71%) than frozen samples (66%) in a four‐class model. Considering the small variation within the first two (NF, LF) and the last two (HF, VF) groups, a two‐class classification model was developed using thawed samples, and the obtained model correctly classified the two groups ([NF, LF] and [HF, VF]) with 100 % classification accuracy. Protein‐ and water‐related changes were found important to distinguish the fillets. Based on these findings, the four‐class prediction model is found insufficient to be used for nondestructive determination of temperature history of frozen fillets. However, the two‐class prediction model with further external validation can be applied to determine the level of temperature fluctuation particularly using fillets scanned at thawed state. Practical Application: NIR spectroscopy can be used to evaluate the degree of temperature fluctuation and thus related quality loss throughout the logistics of frozen Atlantic salmon fillets. Researchers, food control authorities, and the retail industry could be the primary beneficiaries of this research output. [ABSTRACT FROM AUTHOR]

• Bulk optical properties of tissue layers were measured for Satsuma mandarin. • The pigment related μ a and the μ s in juice vesicles increased along harvest time. • Color change was presented in the μ a of flavedo along harvest time. • Influence from fruit size was related to the development stage. • The difference in the μ s of flavedo among cultivars enhanced during maturation. As Satsuma mandarin is a multi-layered fruit, the optical properties of the different tissue layers influence the performance of spectroscopy-based quality detection models. Therefore, the variation in the bulk optical properties (BOP) of the inner (juice vesicles) and outer (flavedo) tissue layer of Satsuma mandarin was investigated for different harvest times, fruit sizes and cultivars. Satsuma mandarin fruit from three different cultivars (Iwasaki, Okitsu and Goku Wase) were harvested at three different times around the recommended harvest date and the BOP of the inner and outer tissue layers were estimated from double integrating spheres measurements. Along the harvest time, the absorption peak related to carotenoids in the bulk absorption coefficient (μ a) spectrum increased in both tissue layers, while the values attributed to chlorophylls decreased in the flavedo. Moreover, a slightly increasing trend over the harvest time was observed in the bulk scattering coefficient (μ s) spectra of juice vesicles, especially at the early stage. The μ a values of both tissue layers and the μ s values of juice vesicles deviated among different fruit sizes, which can be related to the development stage. Smaller fruit presented a higher carotenoids concentration in juice vesicles, an advanced color change in flavedo throughout the maturation stages and higher μ s values in juice vesicles at earlier maturation stages. The μ a differences in pigment absorption among different cultivars were related to the maturation order. Earlier maturing cultivars possessed higher μ a values related to carotenoids absorption in juice vesicles and advanced color changes in flavedo. The μ s values of the flavedo differed among the cultivars and these differences were enhanced during maturation. These insights in the sources of variation in the bulk optical properties of Satsuma mandarin tissues can be used for simulation-based optimal design of an optical measurement configuration and interpretation of the acquired signals in spectroscopy-based quality detection and monitoring. [ABSTRACT FROM AUTHOR]

Lamb’s lettuce ( Valerianella locusta L.) can be stored up to 28 days without being indistinguishable from fresh material by the human eye. However, due to the prior storage period the shelf life potential is limited and this leads to losses in distribution and a lower quality for the consumer. This work aims to develop a rapid and non-destructive methodology using visible/near infrared (Vis/NIR) reflectance spectroscopy to detect and quantify a prior storage period. Vis/NIR reflectance spectra were linked to the time in storage by partial least squares regression (PLS). Different variable selection techniques (interval PLS, Variable Importance in Projection scores, Genetic Algorithms PLS and Monte Carlo Uninformative Variable Elimination PLS) were combined to improve the accuracy and robustness of the prediction model, while decreasing the number of used wavelengths. The final model used only 10% of the original wavelengths, while the root mean squared error of cross validation decreased from 6.0 to 3.6 days. The final model was tested using 2 external test sets and had a maximum root mean square error of prediction of 3.7 days. Vis/NIR reflectance spectroscopy can be a valuable, rapid and non-destructive method for identifying and quantifying a prior storage period of lamb’s lettuce. [ABSTRACT FROM AUTHOR]

Multivariate spectral signals are highly correlated. Often, variable selection techniques are deployed, aiming at model optimization, identification of key variables to explore the underlying physicochemical system or development of a cheap multi-spectral system based on key variables. However, many times the selected variables do not supply a good estimate of properties when tested on a new setting such as new measurements performed on a different spectrometer, different physical or chemical state of the samples and difference in the environmental factors around the experiment. Often the model based on variables selected in the first domain (specific conditions/instrument) does not generalize on the new domain (specific conditions/instrument). To deal with it, in the present work a new method to variable selection called domain invariant covariate selection (di-CovSel) is proposed. The method selects the most informative variables which are invariant to the differences in the instruments, physical or chemical state of the samples and the differences in the environmental factors around the experiment. The method is inspired by domain invariant partial least-square (di-PLS) and the covariate selection (CovSel). The potential of the method is demonstrated on four real cases related to the calibration of near-infrared (NIR) spectroscopy on agri-food materials. The results show that in all the cases, the domain invariant features selected by the di-CovSel have low prediction error compared to the standard variable selection with the CovSel approach when the models are tested on a new data domain. In summary, domain invariant features selected across domains support the development of calibration models with good generalization and supply a better understanding of the system by bypassing the external factors originating from differences in the instruments, physical or chemical states of the samples and the differences in the environmental factors around the experiment. Note that one key feature of the proposed method is that the most important variables which generalize well across domains can be identified without requiring reference measurements in the target domain. • A domain invariant variable selection technique is presented. • Method selects variables to achieve generalized models. • Several application cases for domain invariant variable selection are presented. [ABSTRACT FROM AUTHOR]

Multivariate spectral signals are highly correlated. Often, variable selection techniques are deployed, aiming at model optimization, identification of key variables to explore the underlying physicochemical system or development of a cheap multi-spectral system based on key variables. However, many times the selected variables do not supply a good estimate of properties when tested on a new setting such as new measurements performed on a different spectrometer, different physical or chemical state of the samples and difference in the environmental factors around the experiment. Often the model based on variables selected in the first domain (specific conditions/instrument) does not generalize on the new domain (specific conditions/instrument). To deal with it, in the present work a new method to variable selection called domain invariant covariate selection (di-CovSel) is proposed. The method selects the most informative variables which are invariant to the differences in the instruments, physical or chemical state of the samples and the differences in the environmental factors around the experiment. The method is inspired by domain invariant partial least-square (di-PLS) and the covariate selection (CovSel). The potential of the method is demonstrated on four real cases related to the calibration of near-infrared (NIR) spectroscopy on agri-food materials. The results show that in all the cases, the domain invariant features selected by the di-CovSel have low prediction error compared to the standard variable selection with the CovSel approach when the models are tested on a new data domain. In summary, domain invariant features selected across domains support the development of calibration models with good generalization and supply a better understanding of the system by bypassing the external factors originating from differences in the instruments, physical or chemical states of the samples and the differences in the environmental factors around the experiment. Note that one key feature of the proposed method is that the most important variables which generalize well across domains can be identified without requiring reference measurements in the target domain. • A domain invariant variable selection technique is presented. • Method selects variables to achieve generalized models. • Several application cases for domain invariant variable selection are presented. [ABSTRACT FROM AUTHOR]

Highlights: [•] The optical properties of Braeburn apples before and after shelf-life were measured with a setup for spatially resolved diffuse reflectance spectroscopy. [•] The estimated absorption spectra showed good correlation with the SSC. [•] The estimated absorption spectra gave better prediction of the soluble solids content and firmness than the scattering spectra. [•] The diffuse reflectance spectra gave comparable prediction accuracies to the μ a spectra. [ABSTRACT FROM AUTHOR]

Abstract: The effect of cultivar, season, shelf-life and origin on the accuracy of near infrared (NIR) calibration models for the soluble solids content (SSC) and firmness of apple was studied based on a large spectral data set based on approximately 6000 apple fruit from different cultivars, origins, shelf-life exposure time and seasons. To interpret the variance in the spectra with respect to biological variability, functional analysis of variance (FANOVA) was used. From the FANOVA analysis it was concluded that the effects of cultivar, origin and shelf-life exposure time on the NIR spectra were all significant. The largest differences in the spectra were found around the water absorption peaks (970, 1170 and 1450nm). External validations using independent data sets showed that the accuracy of the models increased considerably when more variability was included in the calibration data set. In general the RMSEP for predictions of the SSC were in the range 0.6–0.8°Brix, while for Magness Taylor firmness it was 5.9–8.8N, depending on the cultivar. It was shown that atypical data can lead to large validation errors. It is, therefore, important to collect a calibration data set which is sufficiently representative for future samples to be analyzed with the developed calibration models and to develop simple procedures for model adaptation during practical use. [Copyright &y& Elsevier]

Abstract: Continuous wave (CW) and time-resolved (TRS) near infrared (NIR) reflectance measurements were carried out on ‘Conference’ pear fruit. While the absorption coefficient spectra obtained from the TRS measurements were dominated by the overtone of the hydroxyl moiety at 975nm, the scattering coefficient spectra were almost flat. The spectral data were used to construct calibration models for the soluble solids content (SSC) and Magness Taylor firmness of ‘Conference’ pear fruit. Reasonable models for SSC (RMSEP=0.44 °Brix) were obtained when the CW spectra in the range 780–1700nm were used, but not when the range was limited to that of the TRS equipment (875–1030nm). No satisfactory calibration model could be established between the TRS absorption coefficient spectrum and SSC. With none of the spectroscopic techniques could significant calibration models for firmness be constructed, although a highly nonlinear relationship between scattering coefficients at 900nm and firmness was observed. [Copyright &y& Elsevier]