•Performance evaluation of designed online diffuse reflectance system.•Stability study of spectra acquired in six fruit orientations using SNR and ACR.•Comparison to SSC models of local and global ...orientations using PLS and LS-SVM.•Selection and discussion of effective wavelengths of global orientation model.
The effect of variation of fruit orientation on online prediction of soluble solids content (SSC) of ‘Fuji’ apples based on visible and near-infrared (Vis/NIR) spectroscopy was studied. The diffuse reflectance spectra in 550–950 nm were collected with a designed online system in six orientations: stem-calyx axis vertical with stem upward (T1) and stem downward (T5), 45° between stem-calyx axis and horizontal with stem slope upward (T2) and stem slope downward (T4), stem-calyx axis horizontal with stem towards computer side lights (T3), stem-calyx axis horizontal with stem towards belt movement direction (T6). The 180 samples with SSC range of 8.00–13.60°Brix were divided into 135 of calibration set with 1.09 standard deviation (S.D.) and 45 of prediction set with 0.85 S.D. The signal-to-noise ratio (SNR) and area change rate (ACR) were used to evaluate the stability of collected spectra. After the comparison of different preprocessing methods, partial least squares (PLS) and least squares-support vector machine (LS-SVM) were used to develop compensation models of SSC for each orientation separately (local models) and all orientations (global model), respectively. Finally, competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA), and their combination were used to select the effective wavelengths (EWs), respectively. Results showed that T1 performed better for our system and influence of measurement orientation on spectra greatly affected SSC prediction accuracy. Comparatively, global model was insensitive to fruit orientation variation. 37 EWs selected by CARS-SPA-PLS model after Savitzky-Golay smoothing in all orientations achieved better results with rp and RMSEP of 0.815, 0.818, 0.837, 0.731, 0.807, 0.842 and 0.487, 0.484, 0.460, 0.573, 0.497, 0.453°Brix, respectively. Generally, global model with EWs could be promisingly used for online SSC prediction of apple.
The soluble solids content (SSC) is an important factor for determining the harvest time and the optimal storage time of apple fruit. However, changes in environmental temperatures cause the apple ...spectrum to fluctuate, which affects the robustness and accuracy of the model for predicting the apple fruit SSC. A portable detection device applicable for different temperatures was developed using a micro-spectrometer and temperature sensor. The detection time for the device was less than 2 s. The circuit board for collecting and analysing the spectrum signal, the detection light path and the housing of the device were designed, and the optimal integration time (6 ms) and light source power (4.5 W) of the device were determined. A total of 420 apple samples were examined in three ambient temperature ranges (0–2, 10–13, and 19–24 °C). The results showed that the model built independently in a single temperature environment can make a good prediction of the samples in the same temperature range. The application of d-value curve method reduced the influence of temperature on model accuracy to a certain extent. The d-value curve is a variable selection method independent of the reference value y, which describes the characterisation ability of each wavelength to samples in different sampling environments. These results conform that removing temperature-sensitive wavelengths from the independent variables or adding temperature factors can both improve the prediction ability of the device in different temperature environments. The model established by incorporating the temperature factor produced the most accurate predictions, with a predicted correlation coefficient, root mean square error of prediction, and ratio of standard deviation to RMSEP of 0.871, 0.402%, and 2.038, respectively.
Display omitted
•A portable optical device was developed to measure the apple SSC.•Temperature compensation improved the prediction accuracy.•Wavelength screening techniques were used to improve the model accuracy.•This device can be used at different environmental temperatures.
•An online spectrum collection system with full transmittance mode was developed.•The features of transmittance spectra collected from different positions were analyzed.•The combinations of different ...spectral excursion correction methods were compared.•The effect of multipoint spectral intensity on prediction accuracy were analyzed.•Variable selection algorithm were used to further eliminate the useless variables.
Higher accuracy for a prediction model is the unremitting pursuit in the field of optics nondestructive technology. The multipoint full-transmittance spectra ranging from 650 to 1000 nm were acquired at a speed of 0.5 m/s using an on-line spectrum measurement system. The combination of mean normalization and 11 points smoothing were selected as the best spectral preprocessing method for removing undesirable signal excursion and light scatters existed in the original spectra. By investigating the interference of transmittance spectral intensity on the prediction accuracy with the method of efficient spectrum optimization proposed in our study, we found that those transmittance spectra with intensity lower than 0.4 at 920 nm were inefficient for SSC prediction. Furthermore, three different variable selection algorithms were used to select characteristic band for further optimizing the prediction model, the best prediction model was built based on 45 variables selected by random flog (RF) and the performance of the best model was Rpre = 0.9043 and RMSEP = 0.4787 respectively. All mentioned above illustrated that efficient spectrum optimization method coupled with variable selection algorithms were useful for improving the accuracy and robust of prediction model.
•Two portable spectrometers were used to predict compositions/properties of sugar beet.•Good predictions of beet moisture, soluble solids and sucrose were obtained.•Important wavelengths were ...identified for moisture, SSC and sucrose content prediction.•Portable spectrometry is useful in production and processing of beets for sugar.
Visible and near-infrared spectroscopy, coupled with partial least squares regression, was used to predict the moisture, soluble solids and sucrose content and mechanical properties of sugar beet. Interactance spectra were acquired from both intact and sliced beets, using two portable spectrometers covering the spectral regions of 400–1100nm and 900–1600nm, respectively. Both visible and short-wave near-infrared (400–1100nm) and near-infrared (900–1600nm) spectrometers gave excellent predictions for the moisture, soluble solids and sucrose content of beet slices with the correlations (rp) of 0.89–0.95 and the standard errors of prediction (SEP) of 0.60–0.85. Lower prediction accuracies were obtained for intact beets, with the rp values of 0.75–0.85 and the SEPs of 0.88–1.23. However, the two spectrometers showed a poor ability of predicting the compressive mechanical properties (i.e., maximum force, area and the slope for the force/displacement curve) of both beet slices and intact beets. Using simple correlation analysis, we also identified wavelengths that had strong correlation with the measured compositions of sugar beets. The portable visible and near-infrared spectrometry is potentially useful for rapid assessment of the moisture, soluble solids and sucrose content of sugar beet at harvest and during postharvest handling and processing.
•Effect of four kiwifruit cultivars on μa and μs' of pulp and peel were investigated.•Peel’s μs' of used two kiwifruit species had different change trends with storage time.•Kiwifruit cultivar ...affected the correlation between pulp’s μa and internal quality.•The relationship between pulp’s μs' and internal quality did not change with cultivars.•The μa had greater potential than μs' in identifying kiwifruit cultivars.
It has been reported that fruit cultivars could be detected by using near-infrared spectroscopy. Therefore, it is doubted that the optical properties (OPs) have differences among different fruit cultivars. However, it is not clear the difference in OPs and whether there is a difference in the relationship between OPs and internal quality. To answer these questions, four kiwifruit cultivars in two species, i.e., ‘Hayward’ and ‘Xuxiang’ of Actinidia deliciosa and ‘Huayou’ and ‘Hongyang’ of Actinidia chinensis, were selected as representative samples to obtain OPs (absorption coefficient μa and reduced scattering coefficient μs') of pulp and peel in storage and 950–1650 nm using an integrating sphere system. Then the relationship between the OPs and internal quality (soluble solids content (SSC), moisture content, and firmness) and microstructure were analyzed. Moreover, the optical parameter with greater potential in classifying kiwifruit cultivars was investigated. The results showed that the changes in μa and μs' with time differed among the four cultivars. Significant differences in the values of μa and μs' were noted among the four cultivars at some given times. For a given species, the investigated two cultivars had an obvious difference in OPs. Below about 1400 nm, the pulp’s μa of ‘Hayward’, ‘Huayou’, and ‘Hongyang’ was positively correlated with SSC and negatively related to moisture content and firmness, and the correlation coefficients changed little with wavelength. However, the correlation coefficients changed greatly for ‘Xuxiang’. Kiwifruit cultivars affect OPs and their relationship with the internal quality. The modeling results of the linear discriminant analysis showed that the μa had greater potential than μs' in identifying kiwifruit cultivars. The study provides useful information for identifying fruit cultivars and internal quality by using near-infrared spectroscopy technology.
•NIR approach was assessed in order to determine the SSC and TA.•Three fruits with different internal structures were utiliszed.•Good results were obtained for the apricot.•Unsatisfactory results ...were obtained for the tomato and passion fruit.•Fruits physical features are limitations to NIR approach and should be considered.
NIR Spectroscopy ability was investigated to assess the fruit structure effect (passion fruit, tomato and apricot) on prediction performance of soluble solids content (SSC) and titratable acidity (TA). Relationships between spectral wavelengths and SSC and TA were evaluated through the application of chemometric techniques based on partial least squares (PLS). Good prediction performance was obtained for apricot with correlation coefficients of 0.93 and 0.95 for SSC and TA and root mean square errors of prediction (RMSEP%) of 3.3% and 14.2%, respectively. For the passion fruit and tomato, the prediction models were not satisfactorily accurate due to the high RMSEP. Results showed that NIR technology can be used to evaluate apricot internal quality, however, it was not appropriate to evaluate internal quality in fruits with thick skin, (passion fruit), and/or heterogeneous internal structure (tomato).
•Local calibration improves model robustness of NIR-based SSC prediction.•A possible explanation is that samples of same level of starch are selected.•Selecting similar samples by PLS scores and ...correlation show equivalent effect.
Nondestructive determination of soluble solids content (SSC) has been used in the fruit industry by using near infrared (NIR) spectroscopy. The robustness of prediction models, which is of great importance in practical application, remains a challenge because of the variability of fruit samples associated with different maturity stages and storage status. Local calibration was investigated in this study as means of improving prediction robustness. As robustness is often reduced by extrapolation, we assessed the robustness by the accuracy of predicting extrapolation samples (samples outside the range of the calibration set). Local calibration was effective in improving the robustness of models compared with global calibration. It is proposed that local calibration optimizes the composition of calibration subset by selecting the samples of same level of starch fractions for each sample to be predicted, and thus provides better robustness due to the homogeneity.
The development of rapid and non-destructive prediction technology for fruit quality after harvest could enhance market competitiveness and profitability. The soluble solids content (SSC) is an ...essential quality index of fruit. This study aims to predict the SSC of crown pear using visible/near-infrared (Vis/NIR) spectroscopy (397–1187 nm) with deep learning model MLP-CNN-TCN. Firstly, the spectral information of crown pears was collected, and Savitzky-Golay (SG) smoothing method and Standard normal variate (SNV) were used to preprocess the spectral data. Secondly, the Multi-layer perceptron (MLP) method was used to reduce the dimension of the preprocessed data, and the reduced data was input into a one-dimensional convolutional neural network (1D-CNN) to extract spectral features. Finally, a Temporal convolutional neural network (TCN) was used to establish a regression model to predict pear SSC. This method was also compared with feedforward neural network (FNN), MLP, 1D-CNN, partial least squares regression (PLSR), and support vector regression (SVR). The MLP-CNN-TCN model obtained better prediction performance, with a prediction determination coefficient (RP2) of 0.956 for SSC. This study demonstrated that the combination of Vis/NIR spectroscopy and MLP-CNN-TCN method could rapidly and non-destructively detect SSC of crown pear, and provide a new regression alternative for the prediction of fruit SSC.
•The SSC of crown pears was predicted by visible/NIR spectroscopy.•A novel deep regression model MLP-CNN-TCN to predict pear SSC was proposed.•Data augmentation using all sampling points was investigated for SSC prediction.•Important wavelengths for SSC prediction of MLP-CNN-TCN was identified by Grad-CAM.
•Rapid non-destructive, online monitoring and grading technologies and instruments for fruit SSC are urgently needed, and could help improve the quality of fruits.•Vis-NIR combined with the ...Savitzky-Golay convolution smoothing, first derivative method, and the SPA were applied to extract and select optimal feature wavelengths with high relevance to SSC. Seven wavelengths were selected as optimal feature wavelengths.•A novel portable NIR diffuse reflectance instrument was designed to evaluate and monitor the SSC of intact apples. The results showed that he portable instrument could be modified to be used to determine the SSC of apples.
The soluble solids content (SSC) is an important internal quality parameter of fruits that is monitored on the fruit market to meet different consumer demands. The aim of this study was to develop a self-made portable near-infrared (NIR) diffuse reflectance instrument to evaluate and monitor the SSC of intact apples. The visible and near-infrared (Vis-NIR) spectra of 118 ‘Fuji’ apples were collected using a Vis-NIR diffuse reflectance spectroscopic measurement system in the spectral range of 450–1100 nm. Savitzky-Golay convolution smoothing and first derivative methods were subsequently applied to eliminate noise interference and baseline shift. Successive projections algorithm (SPA) was performed to extract feature wavelengths from the pretreatment spectra, and the back-propagation artificial neural network (BP-ANN) and multivariate nonlinear regression (MNLR) models were employed to evaluate the predictive ability of the extracted and selected feature wavelengths for the SSC of apples. Seven wavelengths (881, 890, 901, 926, 941, 951, 978 nm) were selected as optimal feature wavelengths from the extracted feature wavelengths (18 wavelengths). The MNLR model (R2 = 0.953, RMSE = 0.391 %) exhibited a high prediction accuracy compared to the BP-ANN model (R2 = 0.865, RMSE = 0.754 %). Based on the results, NIR combined with the MNLR model could be applied in the self-made portable instrument to innovatively monitor the SSC of apples. Finally, a self-made portable NIR diffuse reflectance instrument was designed and developed, and the spectral information of 298 ‘Fuji’ apples was collected using the instrument. According to the results, the MNLR model could well predict the SSC of apples (R2 = 0.871, RMSE = 0.687 %). The overall results also revealed that the developed portable NIR instrument is a promising device for rapid non-destructive monitoring of fruit SSC, thereby meeting the practical application requirement of postharvest commercial processing systems. This instrument could also be beneficial to customers, growers, and producers.
•Prediction of Internal quality and browing disorders in pears with VIS-NIR.•Experiment mimics conditions for VIS-NIR fruit sorter in packinghouses.•Benchmark of regression models for SSC, firmness, ...weight and size in pears.•Benchmark of classification models for browning disorder detection in pears.
This study explores the possibility of predicting the soluble solids content (SSC), firmness and the presence of internal browning disorders in ‘Rocha’ pear (Pyrus communis L.) using a single VIS-NIR spectroscopic measurement in semi-transmittance mode. The spectroscopic measurement setup was developed to mimic real world conditions and takes into account geometry and technical requirements of a commercial fruit sorting optical module. The randomness of the fruit position during the spectra acquisition was simulated by sampling each fruit on four sides. Calibration models for internal quality properties were built using individual and/or average side spectra. The results show that models using the spectrum of each side as an individual sample only under-perform slightly relatively to the models based on spectra averages, which are common in the laboratory but very difficult to implement on an automated grading line. The performance of PLS, SVM and Ridge Regression models was compared for the prediction of SSC and firmness. Multiple types of spectra pre-processing were computed and the best combination of model and pre-processing method identified. The lowest RMSEP results for SSC and firmness were 0.7% (R2 = 0.71) and 7.66 N (R2 = 0.68) respectively, achieved using SVM on data pre-processed with Standard Normal Variate corrected 2nd derivative. For the internal disorder detection (browning), a classification benchmark composed by five different models (PLS-LDA, PCA-Logistic Regression, PCA-Extremely Randomized Trees, Extremely Randomized Trees and SVC) was implemented. PLS-LDA applied to the raw spectra presented the highest sensitivity, 76%. The results confirm that simultaneously achieving viable firmness and SSC predictions and internal disorder detection levels in pears is possible using a single VIS-NIR spectral measurement.