Near-infrared (NIR) spectroscopy as an emerging analytical technique was used for the first time to quantitatively detect the watercore degree and soluble solids content (SSC) in apple. To reduce the ...data processing time and meet the needs of practical application, the variable selection methods including synergy interval (SI), successive projections algorithm (SPA), genetic algorithm (GA) and competitive adaptive reweighted sampling (CARS) were used to identify the characteristic variables and simplify the models. The spectral variables closely related to the apple bioactive components were used for the establishment of the partial least squares (PLS) models. The predictive correlation coefficient (Rp), root mean square error of prediction (RMSEP), and residual predictive deviation (RPD) were used to estimate the performance of the models. The CARS-PLS models displayed the best prediction performance using 600–1000 nm spectra with Rp, RMSEP, and RPD values of 0.9562, 1.340% and 3.720 for apple watercore degree; 0.9808, 0.327 oBx and 4.845 for apple SSC, respectively. These results demonstrate the potential of the NIR transmittance spectroscopy technology for quantitative detection of SSC and watercore degree in apple fruit.
•Novel NIR transmittance spectroscopy quantitatively detected the degree of watercore in apple.•The characteristic spectral variables of apple watercore disease detection were studied.•Multiple variables selection simplified and improved the performance of models.•NIR spectroscopy is advantageous as the fast, non-destructive measurements.
The purpose of this study is to compare the performance of prediction models based either on the bulk optical properties (BOP) or with models based on the diffuse reflection and transmission spectra ...to assess the soluble solids content (SSC) or fruit firmness (FF) of ‘Fuji’ apples. Diffuse reflection and transmission spectra of 240 apples in 500–1000 nm were acquired using a self-constructed single integrating sphere system, from which the absorption coefficient (μa) and reduced scattering coefficient (μ′s) were determined. The relationships of μa, μ′s, and the reflection and transmission spectra with the SSC and FF were analyzed, and detection models were established using partial least squares regression coupled with characteristic wavelength selection. Results showed that neither the SSC nor the FF changed obviously during 80-days storage, and the μa and μ′s showed no monotonic increase or decrease pattern. The SSC prediction model based on μa spectra combined with the baseline correction method performed the best (R2p = 0.79, RMSEP = 0.62), while the model based on μ′s spectra was the best for FF (R2p = 0.67, RMSEP = 0.35 N). Overall, the accuracy of the model based on BOP in predicting internal quality was better than that based on reflection and transmission spectra from the same measurement.
•BOP, reflection and transmission spectra of apples were acquired from the same measurement.•Changes of SSC, FF, reflectance, transmittance and BOP during storage were analyzed.•PLSR models of SSC and FF based on BOP were compared with those based on spectra.•Absorption and scattering properties performed the best for SSC and FF prediction.•Quality evaluation based on BOP was superior to reflection and transmission spectra.
Maturity degree and quality evaluation are important for strawberry harvest, trade, and consumption. Deep learning has been an efficient artificial intelligence tool for food and agro-products. ...Hyperspectral imaging coupled with deep learning was applied to determine the maturity degree and soluble solids content (SSC) of strawberries with four maturity degrees. Hyperspectral image of each strawberry was obtained and preprocessed, and the spectra were extracted from the images. One-dimension residual neural network (1D ResNet) and three-dimension (3D) ResNet were built using 1D spectra and 3D hyperspectral image as inputs for maturity degree evaluation. Good performances were obtained for maturity identification, with the classification accuracy over 84% for both 1D ResNet and 3D ResNet. The corresponding saliency maps showed that the pigments related wavelengths and image regions contributed more to the maturity identification. For SSC determination, 1D ResNet model was also built, with the determination of coefficient (
R
2
) over 0.55 of the training, validation, and testing sets. The saliency maps of 1D ResNet for the SSC determination were also explored. The overall results showed that deep learning could be used to identify strawberry maturity degree and determine SSC. More efforts were needed to explore the use of 3D deep learning methods for the SSC determination. The close results of 1D ResNet and 3D ResNet for classification indicated that more samples might be used to improve the performances of 3D ResNet. The results in this study would help to develop 1D and 3D deep learning models for fruit quality inspection and other researches using hyperspectral imaging, providing efficient analysis approaches of fruit quality inspection using hyperspectral imaging.
•Artificial neural network (ANN) architecture of 9:12:1 can accurately predict soluble solids content (SSC) (R2 = 0.9597).•ANN architecture of 9:11:1 can accurately predict titratable acids content ...(TAC) (R2 = 0.9580).•ANN architecture of 9:11:1 can accurately predict the ratio of soluble solids to titratable acid content (R2 = 0.9658).•The N, P, K, Mg content in fruits contribute relatively largely to the SSC, TAC and SSC/TAC of loquat.
Mineral nutrient elements have an important impact on fruit quality, especially on soluble solids (SSC), titratable acid content (TAC) and the ratio of soluble solids to titratable acid (SSC/TAC) in fruits, which are the most important factors determining the taste and flavor of fruits. In this study, multiple linear regression (MLR) and artificial neural networks (ANN) were used to assess the predictive ability of models to predict SSC, TAC, and SSC/TAC in fruits based on mineral elements in fruits. The results showed that compared with the MLR model (R2 = 0.6772, R2 = 0.5520 and R2 = 0.6025, respectively), the ANNs predicted SSC, TAC, SSC/TAC with higher accuracy and effectiveness (R2 = 0.9597, R2 = 0.9580 and R2 = 0.9658, respectively). These results indicated the ANN is an effective tool with good performance in predicting SSC, TAC, and SSC/TAC of loquat. Meanwhile, we also conducted sensitivity analysis to analyze the relative contribution of mineral nutrients in the fruit to SSC, TAC and SSC/TAC. In terms of relative contribution, the N, P, K, Mg contents in fruits contributed relatively largely to SSC, TAC, and SSC/TAC of loquat.
•Nondestructive measurement of apple SSC was explored using NIR HSI.•Different feature selection methods were used to determine effective wavelengths.•PLS and LS-SVR models for SSC prediction were ...established and compared.
Hyperspectral imaging is a promising technique for nondestructive sensing of multiple quality attributes of apple fruit. This research evaluated and compared different mathematical models to extract effective wavelengths for measurement of apple soluble solids content (SSC) based on near infrared (NIR) hyperspectral imaging over the spectral region of 1000–2500 nm. A total of 160 samples were prepared for the calibration (n = 120) and prediction (n = 40) sets. Competitive adaptive reweighted sampling (CARS), successive projections algorithm (SPA), random frog (RF), and CARS-SPA, CARS-RF combined algorithms were used for extracting effective wavelengths from hyperspectral images of apples, respectively. Based on the selected effective wavelengths, different models were built and compared for predicting SSC of apple using partial least squares (PLS) and least squared support vector regression (LS-SVR). Among all the models, the models based on the ten effective wavelengths selected by CARS-SPA achieved the best results, with Rp, RMSEP of 0.907, 0.479 °Brix for PLS and 0.917, 0.453 °Brix for LS-SVR, respectively. The overall results indicated that CARS-SPA can be used for selecting the effective wavelengths from hyperspectral data. Both PLS and LS-SVR can be applied to develop calibration models to predict apple SSC. Furthermore, the wavelengths selected by CARS-SPA algorithm has a great potential for online detection of apple SSC.
A portable visible and near-infrared (Vis/NIR) device could evaluate and monitor internal qualities of fruit on-tree, as well as during storage conditions after harvest. A portable Vis/NIR device ...which consisted of a commercial spectrometer in the spectral range of 400–1000 nm, an interactance fibre optic probe, a novel switch system, and a microcontroller, was developed and its ability for apple soluble solids content (SSC) prediction was evaluated. A switch system was designed for spectra collection, resulting in the acquisition of three spectra for each measurement of apple fruit, namely the white reference, dark reference, and sample spectrum, which can be used to correct the spectrum of apple fruit dynamically. The results showed that the dynamic correction was more promising than the static correction in which the reference spectra were obtained only once. A model for SSC was built using partial least square (PLS), with the coefficient of determination of prediction (Rp2), the root mean square error of prediction (RMSEP), and the ratio of the standard deviation of the reference destructive SSC to the RMSEP (RPD) of 0.777, 0.561%, and 2.114, respectively. The model was then embedded in the custom software to make it possible for the portable device to predict SSC of apple directly, followed by validation using independent sets. The validation results gave Rp2, RMSEP, and RPD of 0.764, 0.672%, 2.029, respectively for SSC prediction under laboratory conditions, and 0.684, 2.777%, and 0.381, respectively for apples on-tree. The prediction results in the field were improved dramatically using the model built by the field data, with Rp2, RMSEP and RPD of 0.690, 0.604%, and 1.794, respectively. The overall results showed that the developed device had considerable potential to detect the SSC of apple in practical situations.
•A prototype of portable Vis/NIR device for apple SSC detection was developed.•The spectrum of apple fruit was corrected dynamically for each measurement.•The device showed potential to predict SSC of apples in the lab and in the field.
Display omitted
•Two setups with different spectral range were compared in SSC prediction.•Different preprocessing and calibration methods were compared.•Effective wavelengths for SSC prediction of ...tomato were selected and tested.
In this study Vis/NIR spectroscopy was applied to evaluate soluble solids content (SSC) of tomato. A total of 168 tomato samples with five different maturity stages, were measured by two developed systems with the wavelength ranges of 500–930 nm and 900–1400 nm, respectively. The raw spectral data were pre-processed by first derivative and standard normal variate (SNV), respectively, and then the effective wavelengths were selected using competitive adaptive reweighted sampling (CARS) and random frog (RF). Partial least squares (PLS) and least square-support vector machines (LS-SVM) were employed to build the prediction models to evaluate SSC in tomatoes. The prediction results revealed that the best performance was obtained using the PLS model with the optimal wavelengths selected by CARS in the range of 900–1400 nm (Rp = 0.820 and RMSEP = 0.207 °Brix). Meanwhile, this best model yielded desirable results with Rp and RMSEP of 0.830 and 0.316 °Brix, respectively, in 60 samples of the independent set. The method proposed from this study can provide an effective and quick way to predict SSC in tomato.
•Long-term performance of a NIR model for SSC prediction was investigated.•The model after S/B correction eliminated the effect of biological variability.•Effective wavelengths for the SSC prediction ...were determined and validated.
The long-term performance of a near-infrared (NIR) calibration model for soluble solids content (SSC) prediction has been investigated using apples with biological variability collected from 2012 to 2018. The NIR spectrum in the range of 4000–10,000 cm−1 was acquired around equator position for each sample. Partial least squares (PLS) was used to develop calibration model based on the samples harvested in 2012 and 2013. The model was then applied to predict the SSC of samples in five separate data sets collected from 2014 to 2018, resulting in a lower performance with higher RMSEP values in the range of 0.704–1.716%. After applying the slope and bias (S/B) correction method, ten samples were selected from each prediction set and used to adjust the model; the prediction results for five independent prediction sets were improved, with RMSEP values ranging from 0.501% to 0.654%. Subsequently, competitive adaptive reweighted sampling (CARS) and successive projections algorithm (SPA) methods were used to select the most effective wavelengths for the determination of SSC. The calibration model built with 15 wavelengths, combined with the S/B correction method, could replace the full spectral range to detect the SSC of apples over a long period of time, with Rp and RMSEP for five prediction sets being 0.919, 0.937, 0.908, 0.896, 0.924 and 0.592, 0.637, 0.513, 0.523, 0.500%, respectively. Overall, the proposed method in this study could make the model valid and robust over a long time and make the biological variability a negligible interference for SSC prediction, thereby providing potential for SSC prediction in practical application.
•The integration of spectra and textural features were used to predict SSC.•The proposed combined PLS showed excellent ability based on the integration.•Effective variables of spectra and correlation ...improved SSC prediction performance.
The objective of this study was to improve the detection accuracy of soluble solids content (SSC) of apples by integrating spectra and textural features. The spectral data were directly extracted from the region of interest (ROI) of hyperspectral reflectance images of apples over the region of 400–1000nm, while the textural features were obtained by a texture analysis conducted on the ROI images based on grey-level co-occurrence matrix (GLCM). A new regression method called combined partial least square (CPLS) was proposed to analyze the integrations of spectra and different kinds of textural features. In this algorithm, the score matrix matrices of the spectral data and textural features were obtained by PLS analysis separately and then used together for calibration. The prediction results indicated that the CPLS model developed with the integration of spectra and correlation feature achieved promising results and improved SSC predictions compared with the spectral data when used alone. Next, stability competitive adaptive reweighted sampling (SCARS) was conducted to select informative wavelengths for SSC prediction. The CPLS model based on the integration of SCARS selected spectra and correlation gave better results than those with the full wavelength range. The correlation coefficient and root mean square errors of prediction set and validation set were 0.9327 and 0.641%, 0.913 and 0.6656%, respectively. Hence, the integration of spectra and correlation extracted from hyperspectral reflectance images, coupled with CPLS and SCARS methods, showed a considerable potential for the determination of SSC in apples.