João Vitor Ferreira Gonçalves

This study develops and validates an integrated framework that combines UAV multispectral imagery and canopy structural metrics with gridded climatic variables to predict soybean (Glycine max (L.) Merrill) foliar potassium (K) status and grain yield. Field experiments were conducted over three consecutive growing seasons (2022–2023, 2023–2024, and 2024–2025) under different potassium fertilisation strategies and environmental conditions. Machine learning models, particularly the random forest algorithm, were applied to multisource datasets, including UAV-derived canopy structural traits (height and canopy area), spectral indices (NDVI), meteorological variables, and fertilisation information. The foliar K prediction models achieved high accuracy (R² up to 0.85), while the yield prediction models achieved R² values between 0.71 and 0.81. The inclusion of the potassium rate and fertilisation strategy further improved model performance, highlighting the strong influence of potassium supply and fertilisation management on plant physiological responses. Interestingly, compared with those required to stabilise grain yield, foliar potassium saturation occurred at substantially higher K₂O rates, indicating the occurrence of luxury potassium uptake. The association of UAV-derived canopy metrics with this pattern suggests that remote sensing may help detect subtle nutritional dynamics that are not directly reflected in yield responses. Model interpretability using SHAP analysis identified relationships within the analysed dataset that were consistent with physiological expectations, with positive contributions associated with canopy vigour and negative contributions associated with thermal stress. In addition, probabilistic SHAP analysis provided a decision-oriented perspective by quantifying yield probabilities under contrasting potassium management regimes and climate scenarios. Overall, within the experimental conditions studied, the proposed framework enabled a rapid assessment of crop nutritional status, yield prediction, and the evaluation of fertilisation strategies. The integration of UAV data, climatic variables, and machine learning provides an interpretable basis for potassium management and soybean yield forecasting within the experimental conditions studied, while broader transferability requires external validation.

Reliable seed accession identification underpins germplasm conservation, traceability and breeding; however, conventional assays remain destructive, labour-intensive and difficult to scale. Here, visible–near-infrared–shortwave infrared (VIS–NIR–SWIR) hyperspectral imaging (HSI; 449.54–2399.17 nm; 563 bands) was used to classify 32 grain–legume accessions (n = 3200 seeds; 100 seeds per accession), comprising 30 common bean (Phaseolus vulgaris L.) landraces plus two outgroup legumes (Vigna angularis (Willd.) Ohwi & Ohashi and Cajanus cajan (L.) Huth). Each seed was represented by one ROI-averaged spectrum obtained from mean representative pixels within a standardised 10 × 10 pixel window at the centre of each seed. A fixed stratified 70:30 seed-level training:test partition was used, with 70 seeds per accession (n = 2240) reserved for fully independent training and 30 seeds per accession (n = 960) reserved as a fully independent test set. Principal component analysis (PCA) captured 97.42% of the spectral variance in the first three components (PC1 = 63.34%, PC2 = 23.78%, and PC3 = 10.31%). One-versus-rest wavelength association mapping revealed a maximum R² of 0.775 at 461.37 nm, and ReliefF concentrated the strongest reduced-band signal within 449.54–456.30 nm and 577.02–597.54 nm. In the original ReliefF-selected 16-band benchmark, the subspace discriminant reached 68.25% macro-F1 and 68.54% balanced accuracy; after edge-band trimming, the alternative 16-band configuration decreased to 60.67% and 60.94%, respectively. With respect to the full-spectrum sensitivity benchmark, linear discriminant analysis achieved 96.35% balanced accuracy, followed by linear SVM (94.17%). Deep learning trained directly on the full 563-band spectra reached 84.90% test accuracy, 84.47% macro-F1, 86.27% precision and 84.90% recall, with MLP_Wide outperforming the convolutional, recurrent and attention-based alternatives. Overall, under controlled laboratory conditions, this benchmark shows that accession discrimination is driven mainly by visible-domain contrasts in the most compact representations, whereas the full spectral context remains important for the most confusable accessions and for cautious future sensor design. The reduced-band findings should therefore be interpreted as exploratory guidance for sensor design rather than as a validated deployment-ready specification.

Rapid, standardised discrimination of soil taxonomic units remains challenging when relying solely on conventional field descriptions and laboratory analyses, particularly at high sampling densities. This study evaluated whether proximal spectroscopy and hyperspectral imaging can support the classification of Brazilian Soil Classification System (SiBCS) suborders and pedogenetic horizons when surface and subsurface spectra are treated separately. Six intact soil monoliths (0.12 × 1.60 m) were collected in Paraná State, southern Brazil, representing one Organossolo (Ooy), three Latossolos (LVd, LVd1, and LVd2) and two Argissolos (PVAd and PVd). For each monolith, 800 spectra were acquired per sensor with a non-imaging VIS–NIR–SWIR spectroradiometer (350–2500 nm), and 800 spectra per sensor per monolith were extracted from the SWIR hyperspectral images (1200–2450 nm). Principal component analysis (PCA) was used to summarise spectral variability, and supervised classification was performed via k-nearest neighbours, random forest, decision tree and gradient boosting for suborders (10-fold cross-validation), and a neural network was used for within-profile horizon classification. PCA indicated that most of the spectral variance was captured by a dominant axis, with clearer separation among suborders in the SWIR space than in the full VIS–NIR–SWIR range. With respect to suborder classification, subsurface spectra outperformed surface spectra, and SWIR outperformed VIS–NIR–SWIR: the best accuracies were 0.96 for subsurface SWIR (gradient boosting; AUC = 0.99; MCC = 0.95) and 0.89 for surface SWIR (k-nearest neighbours; AUC = 0.98; MCC = 0.87). Within-profile horizon classification via VIS–NIR–SWIR achieved accuracies of 0.84–0.97 with the Neural Network, with most misclassifications occurring between adjacent horizons. Overall, subsurface SWIR information provided the most reliable basis for taxonomic discrimination, whereas horizon classification was feasible but reflected gradual spectral transitions along the profile.

The soybean Glycine max (L.) Merrill is a key crop in Brazil’s agricultural sector and is essential for both domestic food security and international trade. However, water stress severely impacts its productivity. In this study, we examined the physiological and biochemical responses of soybean plants to various water regimes via hyperspectral reflectance (350–2500 nm) and machine learning (ML) models. The plants were subjected to eleven distinct water regimes, ranging from 100% to 0% field capacity, over 14 days. Seventeen key physiological parameters, including chlorophyll, carotenoids, flavonoids, proline, stress markers and water content, and hyperspectral data were measured to capture changes induced by water deficit. Principal component analysis (PCA) revealed significant spectral differences between the water treatments, with the first two principal components explaining 88% of the variance. Hyperspectral indices and reflectance patterns in the visible (VIS), near-infrared (NIR), and shortwave-infrared (SWIR) regions are linked to specific stress markers, such as pigment degradation and osmotic adjustment. Machine learning classifiers, including random forest and gradient boosting, achieved over 95% accuracy in predicting drought-induced stress. Notably, a minimal set of 12 spectral bands (including red-edge and SWIR features) was used to predict both stress levels and biochemical changes with comparable accuracy to traditional laboratory assays. These findings demonstrate that spectroscopy by hyperspectral sensors, when combined with ML techniques, provides a nondestructive, field-deployable solution for early drought detection and precision irrigation in soybean cultivation.

The nutritional and commercial value of lettuce (Lactuca sativa L.) is determined by its foliar pigment and phenolic composition, which varies among cultivars. This study aimed to assess the capacity of hyperspectral and applied multispectral imaging, combined with machine learning algorithms, to predict and map key biochemical traits, such as chloroplastidic pigments (chlorophylls and carotenoids) and extrachloroplastidic pigments (anthocyanins, flavonoids, and phenolic compounds). Eleven cultivars exhibiting contrasting pigmentation profiles were grown under controlled greenhouse conditions, and their chlorophyll a and b, carotenoid, anthocyanin, flavonoid, and total phenolic contents were evaluated. Spectral reflectance data were acquired via a Headwall hyperspectral sensor and a MicaSense multispectral sensor, and the pigment contents were quantified via solvent extraction and a UV microplate reader. We developed predictive models via seven machine learning approaches, with partial least squares regression (PLSR) and random forest (RF) emerging as the most robust algorithms for pigment estimation. Chlorophyll a and b are highly and positively correlated (r > 0.9), which is consistent with their hyperspectral reflectance imaging results. The hyperspectral data consistently outperformed the multispectral data in terms of predictive accuracy (e.g., R² = 0.91 and 0.76 for anthocyanins and flavonoids via RF) and phenolic compounds with R² = 0.79, capturing subtle spectral features linked to biochemical variation. Spatial maps revealed strong genotype-dependent heterogeneity in pigment and phenolic distributions, supporting the potential of this approach for cultivar discrimination and pigment phenotyping. These findings demonstrate that hyperspectral imaging integrated with data-driven modelling offers a powerful, nondestructive framework for the biochemical monitoring of leafy vegetables, supporting breeding, precision agriculture, and food quality assessment.

Soybean (Glycine max (L.) Merrill) is highly sensitive to water deficit, particularly during the vegetative phase, when morphological and metabolic plasticity support continued growth and photosynthetic efficiency. We applied eleven water regimes, from full irrigation (W100) to total water withholding (W0), to plants grown under controlled conditions. After 14 days, we quantified morphophysiological, biochemical, leaf optical, gas exchange, and chlorophyll a fluorescence traits. Drought induces significant reductions in leaf area, biomass, pigment pools, and photosynthetic rates (A, g_s, ΦPSII) while increasing the levels of oxidative stress markers (electrolyte leakage, ROS) and proline accumulation. OJIP transients and JIP test metrics revealed reduced electron-transport efficiency and increased energy dissipation for many parameters under severe stress. Principal component analysis (PCA) clearly separated those treatments. PC1 captured growth and water status variation, whereas PC2 reflected photoprotective adjustments. These data show that progressive drought limits carbon assimilation via coordinated diffusive and biochemical constraints and that the accumulation of proline, phenolics, and lignin is associated with osmotic adjustment, antioxidant buffering, and cell wall reinforcement under stress. The combined use of hyperspectral sensors, gas exchange, chlorophyll fluorescence, and multivariate analyses for phenotyping offers a rapid, nondestructive diagnostic tool for assessing drought severity and the possibility of selecting drought-resistant genotypes and phenotypes in a changing stress environment.

The need to maintain soil health and produce more food worldwide has increased, and soil analysis is essential for its management. Although spectroscopy has emerged as an important tool, it is important to focus primarily on predictive modeling procedures rather than specific protocols. This article aims to contribute to a routine work sequence in a hybrid laboratory that seeks to provide the best data for its users. In this study, 18,730 soil samples from the state of Paraná, Brazil, were analyzed using three different laboratories, sensors and geometries for data acquisition. Thirty soil properties were analyzed, some using different chemical methodologies for comparison purposes. After a spectral reading, two literary protocols were applied, and the final prediction results were observed. We applied cubist models, which were the best for our population. The combination of different spectral analysis systems, with a standardized protocol using LB for the ISS detection of discrepant samples, was shown to significantly improve the accuracy of predictions for 21 of the 30 soil properties analyzed, highlighting the importance of choosing the extraction methodology and improving data quality, which have a significant impact on laboratory analyses, reaffirming spectroscopy as an essential tool for the efficient and sustainable management of soil resources.

Modeling spectral reflectance data using machine learning algorithms presents a promising approach for estimating soil attributes. Nevertheless, a comprehensive investigation of the most effective models, parameters, wavelengths, and data acquisition techniques is essential to ensure optimal predictive accuracy. This work aimed to (a) explore the potential of the soil spectral signature obtained in different spectral bands (VIS-NIR, SWIR, and VIS-NIR-SWIR) and, by using hyperspectral imaging and non-imaging sensors, in the predictive modeling of soil attributes; and (b) analyze the accuracy of different ML models in predicting particle size and soil organic carbon (SOC) applied to the spectral signature of different spectral bands. Six soil monoliths, located in the central north region of Parana, Brazil, were collected and scanned via hyperspectral cameras (VIS-NIR camera and SWIR camera) and spectroradiometer (VIS-NIR-SWIR) in the laboratory. The spectral signature of the soils was analyzed and subsequently applied to ML models to predict particle size and SOC. Each set of data obtained by the different sensors was evaluated separately. The algorithms used were k-nearest neighbors (KNN), support vector machine (SVM), random forest (RF), linear regression (LR), artificial neural network (NN), and partial least square regression (PLSR). The most promising predictive performance was observed for the complete VIS-NIR-SWIR spectrum, followed by SWIR and VIS-NIR. Meanwhile, KNN, RF, and NN models were the most promising algorithms in estimating soil attributes for the dataset obtained from both sensors. The general mean R² (determination coefficient) values obtained using these models, considering the different spectral bands evaluated, were around 0.99, 0.98, and 0.97 for sand prediction, and around 0.99, 0.98, and 0.96 for clay prediction. The lower performances, obtained for the datasets from both sensors, were observed for silt and SOC, with R² results between 0.40 and 0.59 for these models. KNN demonstrated the best predictive performance. Integrating effective ML models with robust sample databases, obtained by advanced hyperspectral imaging and spectroradiometers, can enhance the accuracy and efficiency of soil attribute prediction.