Search

• Providing a novel method to detect determine apple valsa canker (AVC) at different infection stages using molecular vibrational spectroscopy information.Selecting several variables based on x -loadings and CARS algorithm. • Developing LS-SVM models for diagnosing AVC infection stages and using ROC to evaluate models. Detecting of apple valsa canker (AVC), which is caused by Valsa mali Miyabe et Yamada, at early infection stage is beneficial to disease prevention and control for ensuring yield and quality of apples. Applying simple, economical and non-destructive rapid detection method for the early diagnosis of AVC is of great significance for precision management of orchards. In this research, near-infrared (NIR) spectroscopy (900−1700 nm) and Raman scattering (0−2000 cm−1) combined with machine learning algorithms were employed to diagnose 3 infection degrees of AVC (healthy, disease 1, disease 2) based on optimal variables which were selected by chemometric methods. NIR and Raman spectroscopy were obtained using portable spectrometers in reflection mode, respectively. Firstly, adaptive iterative reweighting partial least squares (air-PLS) was utilized to remove fluorescence background in Raman spectra. Secondly, clustering analysis was developed using principal component analysis (PCA). After that, optimal variables were selected by x -loadings (XLs) of PCA and competitive adaptive reweighted sampling (CARS) algorithm, respectively. Four optimal wavelengths at 983, 1156, 1395, and 1457 nm from NIR spectra and five optimal wavenumbers at 175, 229, 326, 409, and 1523 cm−1 from Raman spectra were identified by XLs. Six optimal wavelengths at 943, 949, 967, 1240, 1632, 1666 nm from NIR spectra and eight optimal wavelengths at 412, 1500, 1585, 1596, 1599, 1602, 1672 and 1709 cm−1 from Raman spectra were selected according to CARS. Finally, AVC diagnosing models were developed using least square support vector machine (LS-SVM), and receiver operating characteristic (ROC) curves were applied to evaluate the performance of the classification models. Overall, the relatively good classification results on optimal variables (94.67% and 97.33% in NIR, 89.33% and 89.33% in Raman) were obtained using LS-SVM. This study confirmed that molecular vibrational spectroscopy techniques (NIR and Raman) is promising for early detecting AVC and providing a practical way for diagnosing diseases in large-scale orchards. [ABSTRACT FROM AUTHOR]

Detection and characterization of interactions between crop plants and hydrogen peroxide (H 2 O 2) is significant for the exploration of the mechanisms in plant pathology. The objective of this research is to estimate spectral characteristics of rapeseed leaves (Brassica napus L.) during treatment with different H 2 O 2 concentrations (0, 0.5, 1.0, and 3.0 mmol/L) by using Raman spectroscopy (RS) (800–1800 cm−1) and hyperspectral imaging (HSI) (400–1000 nm). Cluster analysis of RS and HSI data between the control and treated samples was conducted using kernel principal component analysis (KPCA) and principal component analysis (PCA), respectively. Characteristic Raman shifts at 1012, 1163, and 1530 cm−1 and hyperspectral featured wavelengths at 452, 558, 655, and 703 nm were selected for discriminating control and treated samples. The one-way analysis of variance (ANOVA) was applied to demonstrate the significant difference in spectral signatures of samples, and results showed that 452 nm is promising to assess the control and treated samples at the p < 0.05 level. The featured Raman shifts and hyperspectral wavelengths were employed to establish least squares-support vector machine (LS-SVM) discriminative models. The approach of multiple-level data fusion of 1163 cm−1 combined with 452 nm produced the best recognize rate (RR) of 81.7% to detect the control and treated leaves than other models. Therefore, the results encouraged multiple sensor fusion to improve models for better model performance and to detect plant treatment situations with H 2 O 2 solutions. Unlabelled Image • Providing a method to determine hydrogen peroxide treatment of plant leaves using Raman spectroscopy and hyperspectral information. • KPCA and PCA provided good cluster results based on Raman shift and hyperspectral data between controlled and treated samples. • Multiple level data fusion of 1163 cm-1 combined with 452 nm produced the better recognize rate of 81.7%. [ABSTRACT FROM AUTHOR]

• Assessed was the blueberry buds damage due to lab-simulated freezing events. • Buds were imaged at bud swell and early pink stages in 2017 and 2018 season. • Salient wavelengths were selected using principal component analysis and successive projection algorithm technique. • Partial least squares discriminant analysis classifier evaluated the robustness of salient wavelengths. The floral tissue of blueberry buds is very sensitive to the freezing and can be damaged during early spring freeze events. Such damage results in loss of yield to the commercial blueberry producers. Therefore, this study investigated the feasibility of injured blueberry buds damage detection during early spring freeze using hyperspectral imaging (HSI) technique. Imaged were the buds from bud swell and early pink stage for two growing seasons. Prior to imaging, buds in 2017 were subjected to freezing temperatures of −7 to −16 °C and −2 to −11 °C at respective growth stages. In 2018 season, buds were frozen at −11 to −20 °C and −4 to −13 °C at two respective sampling date for bud swell stage and at −2 to −11 °C for early pink growth stage. The HSI data was obtained using a push broom type system in the spectral wavelength ranges of 517–1729 nm. For the imaged buds, each of the indivisual buds on sampled shoot was considered as one sample, and the average spectra of such bud within 540 to 1599 nm was extracted for further analysis. Principal component analysis and successive projection algorithms were applied to select key wavelengths related to bud damage symptoms. Common wavelengths of 615, 673, 690, 756, 979 and 1467 nm were identified for both the growth stages. Partial least squares discriminant analysis was performed to identify normal and injured bud samples with the selected common wavelengths. Performance of the models manifested good results with better sensitivity (>0.80), specificity (>0.75), the area under the receiver operating characteristic curve (>0.84) and accuracy (>0.75) attributes for 2017, 2018 seasons and two seasons combined test dataset. [ABSTRACT FROM AUTHOR]

Abstract: Visible–near infrared reflectance (NIR) spectroscopy was applied to identify the subtle bruise of fresh jujube. In this paper, 70 samples were selected randomly as a calibration set and 30 samples were selected as a prediction set. Multiplicative scatter calibration (MSC) was used to pre-process the spectra, 9 characteristic wavelengths were selected by successive projection algorithm (SPA) in the end. On the basis of spectroscopy theoretical analysis, the principal component analysis (PCA) was performed to analyze the spectral data of the characteristic wavelengths, and 4 principal components (PCs) data which were used as the input variables for the Least squares support vector machine (LS–SVM) model for identifying the subtle bruise of fresh Jujube fruits were picked up. Intact jujubes and bruise jujubes in prediction with code 1 and 2 respectively could be considered as the outputs. The identification accuracy rate of the MSC–SPA–PCA–LS–SVM model for subtle bruises of fresh jujube fruits could reach 100%, proving that the key information in spectral data could be effectively picked up by SPA and PCA combination methods. It is feasible to identify the subtle bruise of fresh jujube using visible–near infrared reflectance (NIR) spectroscopy. [Copyright &y& Elsevier]