Search Results (136)

Rapid and accurate moisture detection is critical for alfalfa quality control, yet conventional methods are slow, and non-destructive techniques are challenged by different product forms. This study leveraged Short-Wave Infrared Hyperspectral Imaging (SWIR-HSI) to acquire spatially representative spectra, aiming to develop and validate robust, form-specific moisture prediction models for compressed and powdered alfalfa. For compressed alfalfa, a full-spectrum Support Vector Regression (SVR) model demonstrated stable and good performance (mean Prediction Coefficient of Determination ${\mathrm{R}}_{\mathrm{P}}^2$ = 0.880, Ratio of Performance to Deviation RPD = 2.93). In contrast, powdered alfalfa achieved superior accuracy (mean ${\mathrm{R}}_{\mathrm{P}}^2$ = 0.953, RPD = 5.29) using an optimized pipeline of Savitzky–Golay’s first derivative, Successive Projections Algorithm (SPA) for feature selection, and an SVR model. A key finding is that the optimal model for powdered alfalfa frequently converged to an ultra-sparse, single-band solution near water absorption shoulders (~970/1450 nm), highlighting significant potential for developing low-cost, filter-based agricultural sensors. While this minimalist model showed excellent average accuracy, rigorous repeated evaluations also revealed non-negligible performance variability across different data splits—a crucial consideration for practical deployment. Our findings underscore that tailoring models to specific product forms and explicitly quantifying their robustness is essential for reliable NIR sensing in agriculture and provides concrete wavelength targets for sensor development. Full article

This study addresses the observation of floating plastic debris in freshwater environments using an Unmanned Aerial Vehicle (UAV) multi-sensor strategy. An experimental campaign is described where an heterogeneous plastic assemblage, namely a plastic target, and a naturally occurring leaf-litter mat are observed by a UAV platform in the Lake Calore (Avellino, Southern Italy) within the framework of the “multi-layEr approaCh to detect and analyze cOastal aggregation of MAcRo-plastic littEr” (ECOMARE) Italian Ministry of Research (MUR)-funded project. Three UAV platforms, equipped with optical, multispectral, and thermal sensors, are adopted, which overpass the two targets with the objective of analyzing the sensitivity of optical radiation to plastic and the possibility of discriminating the plastic target from the natural one. Georeferenced orthomosaics are generated across the visible, multispectral (Green, Red, Red Edge, Near-Infrared—NIR), and thermal bands. Two novel indices, the Plastic Detection Index (PDI) and the Heterogeneity Plastic Index (HPI), are proposed to discriminate between the detection of plastic litter and natural targets. The experimental results highlight that plastics exhibit heterogeneous spectral and thermal responses, whereas natural debris showed more homogeneous signatures. Green and Red bands outperform NIR for plastic detection under freshwater conditions, while thermal imagery reveals distinct emissivity variations among plastic items. This outcome is mainly explained by the strong NIR absorption of water, the wetting of plastic surfaces, and the lower sensitivity of the Mavic 3′s NIR sensor under high-irradiance conditions. The integration of optical, multispectral, and thermal data demonstrate the robustness of UAV-based approaches for distinguishing anthropogenic litter from natural materials. Overall, the findings underscore the potential of UAV-mounted remote sensing as a cost-effective and scalable tool for the high-resolution monitoring of plastic pollution over inland waters. Full article

In the framework of the SURPRISE EU project, the Compressive Sensing paradigm was applied for the development of a laboratory demonstrator with improved spatial sampling operating from visible up to Medium InfraRed (MIR). The demonstrator, which utilizes a commercial Digital Micromirror Device modified by replacing its front window with one transparent up to MIR, has 10 bands in the VIS-NIR range and 2 bands in the MIR range, showing a super resolution factor up to 32. Measurements performed in the MIR spectral range using hot sources as targets show that CS is effective in reconstructing super-resolved hot targets. Full article

Dendrobium officinale is a valuable Chinese medicinal herb, but distinguishing it from other Dendrobium species after processing is challenging, leading to low classification accuracy and time-consuming analysis. This study proposes a rapid classification model based on near-infrared hyperspectral imaging (NIR-HSI), incorporating data preprocessing and feature wavelength selection. Five Dendrobium species—D. officinale, D. aphyllum, D. chrysanthum, D. fimbriatum, and D. thyrsiflorum—were used. Spectral preprocessing techniques like normalization and smoothing were applied, and Support Vector Machine (SVM) models were constructed. Normalization improved both accuracy and stability, with the full-spectrum Normalize-SVM model achieving 97% accuracy for calibration and 88% for prediction. D. chrysotoxum performed best, with all metrics reaching 100%, while D. aphyllum had poor classification (40% recall and 51.74% F1 score). To improve efficiency and performance, feature wavelength selection was performed using Competitive Adaptive Reweighted Sampling (CARS) and Successive Projections Algorithm (SPA). The CARS-Normalize-SVM model yielded the best results: 98% accuracy for calibration and 96% for prediction, improving by 1% and 8%, respectively. D. aphyllum’s classification also improved significantly, with a 100% recall rate and 95.24% F1 score. These findings highlight hyperspectral imaging’s potential for rapid Dendrobium species identification, supporting future quality control and market supervision. Full article

Soil organic carbon (SOC) is a critical indicator of soil health and carbon sequestration potential. Accurate, efficient, and scalable SOC estimation is essential for sustainable orchard management and climate-resilient agriculture. However, traditional visible–near-infrared (Vis–NIR) spectroscopy often suffers from limited chemical specificity and weak adaptability in heterogeneous soil environments. To overcome these limitations, this study develops a five-stage modeling framework that systematically integrates Fourier Transform Infrared (FTIR) spectroscopy with hybrid machine learning techniques for non-destructive SOC prediction in citrus orchard soils. The proposed framework includes (1) FTIR spectral acquisition; (2) a comparative evaluation of nine spectral preprocessing techniques; (3) dimensionality reduction via three representative feature selection algorithms, namely the Successive Projections Algorithm (SPA), Competitive Adaptive Reweighted Sampling (CARS), and Principal Component Analysis (PCA); (4) regression modeling using six machine learning algorithms, namely the Random Forest (RF), Support Vector Regression (SVR), Gray Wolf Optimized SVR (SVR-GWO), Partial Least Squares Regression (PLSR), Principal Component Regression (PCR), and the Back-propagation Neural Network (BPNN); and (5) comprehensive performance assessments and the identification of the optimal modeling pathway. The results showed that second-derivative (SD) preprocessing significantly enhanced the spectral signal-to-noise ratio. Among feature selection methods, the SPA reduced over 300 spectral bands to 10 informative wavelengths, enabling efficient modeling with minimal information loss. The SD + SPA + RF pipeline achieved the highest prediction performance (R² = 0.84, RMSE = 4.67 g/kg, and RPD = 2.51), outperforming the PLSR and BPNN models. This study presents a reproducible and scalable FTIR-based modeling strategy for SOC estimation in orchard soils. Its adaptive preprocessing, effective variable selection, and ensemble learning integration offer a robust solution for real-time, cost-effective, and transferable carbon monitoring, advancing precision soil sensing in orchard ecosystems. Full article

This study addresses the urgent need for the rapid, non-destructive assessment of key soybean components, including moisture, fat, and protein, using near-infrared (NIR) spectroscopy. This study provides technical and theoretical support for achieving the efficient and accurate detection of major soybean components and for the development of portable near-infrared (NIR) instruments. Thirty soybean samples from diverse sources were collected, and 360 spectral measurements were acquired using a 900–1700 nm NIR spectrometer after grinding and standardized sampling. To improve model robustness, preprocessing strategies such as standard normal variate (SNV), multiplicative scatter correction (MSC), and Savitzky–Golay derivatives were applied. Feature selection was conducted using competitive adaptive reweighted sampling (CARS), successive projections algorithm (SPA), and uninformative variable elimination (UVE), followed by model construction with partial least squares regression (PLSR), support vector regression (SVR), and random forest (RF). Comparative analysis revealed that the RF model consistently outperformed the others across most combinations. Specifically, the SPASNV + D₁–RF combination achieved an RPD of 14.7 for moisture, CARS–SNV + D₁–RF reached 5.9 for protein, and CARS–SG + D₂–RF attained 12.0 for fat, all significantly surpassing alternative methods and demonstrating a strong nonlinear learning capacity and predictive precision. These findings show that integrating optimal preprocessing and feature selection strategies can markedly enhance the predictive accuracy in NIR-based soybean analyses. The RF model offers exceptional stability and performance, providing both technical reference and theoretical support for the development of portable NIR devices and practical rapid-quality assessment systems for soybeans in industrial applications. Full article

The objective of this study was to evaluate the feasibility of using visible, near-infrared, and short-wave infrared (VIS/NIR/SWIR) spectroscopy coupled with chemometrics for non-destructive prediction of nutritional components in Galia-type melon fruit. A total of 175 fully ripened melons were analyzed for soluble solids content (SSC), dry matter (DM), pH, and titratable acidity (TA) using partial least squares regression (PLSR), principal components regression (PCR), and multilinear regression (MLR) models. Reflectance spectra were captured at three fruit locations (pedicel, equatorial, and blossom end) in the 350–2500 nm range. The PLSR models yielded the highest accuracy, particularly for SSC (R = 0.80) and SSC/TA (R = 0.79), using equatorial zone data. Variable selection using the genetic algorithm (GA) successfully identified the spectral regions critical for each nutritional parameter at the pedicel, equatorial, and blossom end areas. Key wavelengths for SSC were found around 670–720 nm and 900–1100 nm, with important wavelengths for pH prediction located near 1450 nm, and, for dry matter, in the ranges 1900–1950 nm. Variable importance in projection (VIP) analysis confirmed that specific wavelengths between 680 and 720 nm, 900 and 1000 nm, 1400 and 1500 nm, and 1900 and 2000 nm were consistently critical in predicting the SSC, DM, and SSC/TA ratio. The highest VIP scores for SSC prediction were noted around 690 nm and 950 nm, while dry matter prediction was influenced most by wavelengths in the 1450 nm to 1950 nm range. This study demonstrates the potential of VIS/NIR/SWIR spectroscopy for rapid, non-destructive melon quality assessment, with implications for commercial postharvest management. Full article

Fire is a destructive hazard impacting residential, industrial, and forested environments. Once ignited, fire becomes difficult to control, and recovery efforts are often extensive. Therefore, early detection is critical for effective firefighting. This study presents a mobile robotic system designed for early fire detection, integrating a Raspberry Pi, RGB (red, green and blue), and night vision-NIR (near infrared reflectance) cameras. A four-stage hybrid-cascade machine learning model was developed by combining state-of-the-art (SotA) models separately trained on RGB and NIR images. The system accounts for both daytime and nighttime conditions, achieving F1 scores of 96.7% and 95.9%, respectively, on labeled fire/non-fire datasets. Unlike previous single-stage or two-stage vision pipelines, our work delivers a lightweight four-stage hybrid cascade that jointly fuses RGB and NIR imagery, integrates temporal consistency via ConvLSTM, and projects a robot-centric “safe-approach distance” in real time, establishing a novel edge-level solution for mobile robotic fire detection. Based on real-life test results, the robotic system with this new hybrid-cascade model could detect the fire source from a safe distance of 500 mm and with notably higher accuracy compared to structures with other models. Full article

In this investigation, we established an intelligent computational framework comprising a novel starfish-optimization-algorithm-optimized support vector regression (SOA-SVR) model and a multi-algorithm joint strategy to evaluate the processing applicability of wheat flour in terms of sedimentation value (SV) and falling number (FN) using near-infrared (NIR) data (900–1700 nm) obtained using a miniaturized NIR spectrometer. By employing an improved whale optimization algorithm (iWOA) coupled with a successive projections algorithm (SPA), we selected the 20 most informative wavelengths (MIWs) from the full range spectra, allowing the iWOA/SPA-SOA-SVR model to predict SV with correlation coefficient and root-mean-square errors in prediction (R_P and RMSE_P) of 0.9605 and 0.2681 mL. Additionally, RFE, in combination with the iWOA, identified 30 MIWs and enabled the RFE/iWOA-SOA-SVR model to predict the FN with an R_P and RMSE_P of 0.9224 and 0.3615 s. The robustness and reliability of the two SOA-SVR models were further validated using 50 independent samples per index, a statistical two-sample F-test, and a t-test. In conclusion, the combination of a portable miniaturized NIR spectrometer and an SOA-driven SVR algorithm demonstrated technical feasibility in quantifying the SV and FN of wheat flour, thus providing a novel strategy for the on-site assessment of wheat flour processing applicability. Full article

The difficulty in controlling the quality of Korla pears is the main factor limiting their market value. The key to solving this problem is to detect the color of Korla pears quickly and accurately. This study employed near-infrared spectroscopy (NIRS) technology to measure the absorbance of Korla fragrant pears. The full-spectrum data were pre-processed using six methods: Savitzky–Golay convolution smoothing (SGCS), Savitzky–Golay convolution derivative (SGCD), multiplicative scatter correction (MSC), vector normalization (VN), min–max normalization (MMN), and standard normal variate transformation (SNV). The pre-processed spectral data were subjected to characteristic band extraction using the successive projections algorithm (SPA) and uninformative variable elimination (UVE) methods. Subsequently, detection models for the color indices L*, a*, and b* of Korla fragrant pears were established using the partial least squares regression (PLSR) with full-spectrum and characteristic extracted spectral data. The optimal detection models were determined. The results indicated that pre-processing and characteristic extraction improved the accuracy of the PLSR model. The optimal detection model for the color index L* was SGCD-UVE-PLSR (correlation coefficient (R) = 0.80, Root Mean Square Error (RMSE) = 1.19); for the color index a*, it was VN-SPA-PLSR (R = 0.84 and RMSE = 1.28), and for the color index b*, it was MSC-UVE-PLSR (R = 0.84 and RMSE = 1.25). This research provides a theoretical reference for developing color detection instruments for Korla fragrant pears. Full article

The act of performing music may induce a specific state of mind, musicians potentially becoming immersed and detached from the rest of the world. May this be measured? Does this state of mind change based on repetition? In collaboration with Stavanger Symphony Orchestra (SSO), we developed protocols to investigate ongoing changes in the brain activation of a first violinist and a second violinist in real time during seven sequential, public concerts using functional near-infrared spectroscopy (fNIRS). Using wireless fNIRS systems (Brite MKII) from Artinis, we measured ongoing hemodynamic changes and projected the brain activation to the audience through the software OxySoft 3.5.15.2. We subsequently developed protocols for further analyses through the Matlab toolboxes Brainstorm and Homer2/Homer3. Our developed protocols demonstrate how one may use “functional dissection” to imply how the state of mind of musicians may alter while performing their art. We focused on a subset of cortical regions in the right hemisphere, but the current study demonstrates how fNIRS may be used to shed light on brain dynamics related to producing art in ecological and natural contexts on a general level, neither restricted to the use of musical instrument nor art form. Full article

This study evaluates the potential of integrating visible and near-infrared (visNIR) spectroscopy, color analysis, and firmness testing for non-destructive tomato quality assessment and shelf-life prediction. Tomato fruit (cv. HM1823) harvested at four ripening stages were monitored over 12 days at 22 °C to investigate ripening stage-specific variations in key quality parameters, including color (hue angle), firmness (compression), and nutritional composition (pH, soluble solids content, and titratable acidity ratio). Significant changes in these parameters during storage highlighted the need for advanced tools to monitor and predict quality attributes. Spectral data (340–2500 nm) captured using advanced and cost-effective portable spectroradiometers, coupled with chemometric models such as partial least squares regression (PLSR), demonstrated reliable predictions of shelf-life and nutritional quality. The near-infrared spectrum (900–1700 nm) was particularly effective, with variable selection methods such as genetic algorithm (GA) and variable importance in projection (VIP) scores enhancing model accuracy. This study highlights the promising role of visNIR spectroscopy as a rapid, non-destructive tool for optimizing postharvest management in tomato. By enabling real-time quality assessments, these technologies support sustainable agricultural practices through improved decision-making, reduced postharvest losses, and enhanced consumer satisfaction. The findings also validate the utility of affordable spectroradiometers, offering practical solutions for stakeholders aiming to balance cost efficiency and reliability in postharvest quality monitoring. Full article

Rapid and accurate detection of protein content is essential for ensuring the quality of maize. Near-infrared spectroscopy (NIR) technology faces limitations due to surface effects and sample homogeneity issues when measuring the protein content of whole maize grains. Focusing on maize grain powder can significantly improve the quality of data and the accuracy of model predictions. This study aims to explore a rapid detection method for protein content in maize grain powder based on near-infrared spectroscopy. A method for determining protein content in maize grain powder was established using near-infrared (NIR) reflectance spectra in the 940–1660 nm range. Various preprocessing techniques, including Savitzky−Golay (S−G), multiplicative scatter correction (MSC), standard normal variate (SNV), and the first derivative (1D), were employed to preprocess the raw spectral data. Near-infrared spectral data from different varieties of maize grain powder were collected, and quantitative analysis of protein content was conducted using Partial Least Squares Regression (PLSR), Support Vector Machine (SVM), and Extreme Learning Machine (ELM) models. Feature wavelengths were selected to enhance model accuracy further using the Successive Projections Algorithm (SPA) and Uninformative Variable Elimination (UVE). Experimental results indicated that the PLSR model, preprocessed with 1D + MSC, yielded the best performance, achieving a root mean square error of prediction (RMSEP) of 0.3 g/kg, a correlation coefficient (R_p) of 0.93, and a residual predictive deviation (RPD) of 3. The associated methods and theoretical foundation provide a scientific basis for the quality control and processing of maize. Full article

In the petroleum industry, accurately identifying downhole fluid is crucial for understanding fluid composition and estimating crude oil contamination and other properties. Near-infrared (NIR) spectrum analysis technology has achieved successful fluid identification applications due to its non-destructive nature and high efficiency. However, for real-time downhole fluid analysis, the NIR spectrometer faces challenges such as miniaturization and cost effectiveness. To address these issues, we construct a real-time downhole fluid identification system in this work. First, we propose a lightweight and deployable fluid identification model by integrating the successive projections algorithm (SPA) and a quadratic neural network (QNN). The SPA allows for spectral feature selection, and the QNN acts as an efficient identification model. Consequently, we use only four specific wavelengths with a one-hidden-layer QNN to achieve high identification accuracy. Second, we devise a hardware deployment scheme for real-time identification. We use four laser diodes to replace conventional light sources, further saving space. The QNN is then deployed to the STM32 MCU to implement real-time identification. Computational and online experiments demonstrate that our system functions well in real-time fluid identification and can further estimate the oil contamination rate with acceptable error. Full article

The market demand for baby leaf lettuce is constantly increasing, while safety has become one of the most important traits in determining consumer preference driven by human health hazards concerns. In this study, the performance of visible and near-infrared (vis/NIR) spectroscopy was tested in discriminating pesticide-free against pesticide-treated lettuce plants. Two commercial fungicides (mancozeb and fosetyl-al) and two insecticides (deltamethrin and imidacloprid) were applied as spray solutions at the recommended rates on baby leaf lettuce plants. Untreated-control plants were sprayed with water. Reflectance data in the wavelength range 400–2500 nm were captured on leaf samples until harvest on the 10th day upon pesticide application, as well as after 4 and 8 days during post-harvest storage at 5 °C. In addition, biochemical components in leaf tissue were also determined during storage, such as antioxidant enzymes’ activities (peroxidase [POD], catalase [CAT], and ascorbate peroxidase [APX]), along with malondialdehyde [MDA] and hydrogen peroxide [H₂O₂] content. Partial least square discriminant analysis (PLSDA) combined with feature-selection techniques was implemented, in order to classify baby lettuce tissue into pesticide-free or pesticide-treated ones. The genetic algorithm (GA) and the variable importance in projection (VIP) scores identified eleven distinct regions and nine specific wavelengths that exhibited the most significant effect in the detection models, with most of them in the near-infrared region of the electromagnetic spectrum. According to the results, the classification accuracy of discriminating pesticide-treated against non-treated lettuce leaves ranged from 94% to 99% in both pre-harvest and post-harvest periods. Although there were no significant differences in enzyme activities or H₂O₂, the MDA content in pesticide-treated tissue was greater than in untreated ones, implying that the chemical spray application probably induced a stress response in the plant that was disclosed with the reflected energy. In conclusion, vis/NIR spectroscopy appears as a promising, reliable, rapid, and non-destructive tool in distinguishing pesticide-free from pesticide-treated lettuce products. Full article