Agricultural Engineering

Abstract Introduction Soil is a fundamental non-living component of ecosystems and plays a crucial role in agricultural production, natural resource sustainability, and environmental protection. As a medium for plant growth, nutrient cycling, and biological activity, healthy soil forms the basis of food security and ecosystem stability. Among the various forms of land degradation, soil salinization is considered one of the most severe environmental challenges, particularly in arid and semi-arid regions. Soil salinity significantly reduces soil fertility, limits plant growth, decreases biological productivity, and threatens the sustainability of rangeland ecosystems. Given the extensive distribution of arid and semi-arid lands in Iran, accurate monitoring and mapping of soil salinity are essential for sustainable land management. Conventional field-based methods, although accurate, are labor-intensive, expensive, and incapable of providing adequate spatial coverage over large areas. Furthermore, field sampling offers only point-based information and a limited representation of salinity conditions. In contrast, remote sensing technologies integrated with machine learning techniques provide efficient and cost-effective tools for large-scale soil salinity assessment and mapping. The present study aimed to evaluate the effectiveness of remote sensing-derived variables and different machine learning algorithms for soil salinity classification and mapping in the rangelands of Ramhormoz County, located in Khuzestan Province, southwestern Iran.
Materials and Methods A total of 80 surface soil samples were randomly collected from the rangelands of Ramhormoz County at a depth of 0–30 cm, with a relatively uniform spatial distribution across the study area. The geographic coordinates of all sampling locations were recorded using a Global Positioning System (GPS) receiver to ensure accurate correspondence between field observations and satellite image pixels. Soil salinity was determined by measuring the electrical conductivity of the saturated paste extract (ECe) in the laboratory. Landsat 9 imagery acquired in April 2024, close to the sampling period, was processed using Google Earth Engine (GEE). Spectral bands, topographic information (elevation), and several salinity-related spectral indices, including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Brightness Index (BI), Salinity Index (SI), Salinity Index 1 (SI1), Salinity Index 2 (SI2), and Normalized Difference Salinity Index (NDSI), were extracted and used as predictor variables. To reduce multicollinearity among input variables, Pearson correlation analysis was conducted. Variables exhibiting strong pairwise correlations (r > 0.90) were excluded from subsequent analyses. Soil salinity classification was performed using four machine learning algorithms: Classification and Regression Trees (CART), Random Forest (RF), Support Vector Machine (SVM), and Multilayer Perceptron Neural Network (MLP). The performance of the models was evaluated using 10-fold cross-validation (K = 10), overall accuracy, weighted F1-score, and the coefficient of determination (R²).
Results and Discussion The measured soil salinity values ranged from 1.79 to 11.30 dS m⁻¹, indicating moderate variability in salinity conditions and the absence of extremely saline soils within the study area. This range provided an appropriate dataset for training and evaluating the machine learning models. Comparative analysis of the tested algorithms revealed notable differences in predictive performance. The CART model demonstrated the highest classification accuracy, achieving an overall accuracy of 75% and a weighted F1-score of 74.8%, indicating its superior capability in distinguishing salinity classes. The SVM model with a linear kernel ranked second, with an overall accuracy of 63%. In contrast, the Random Forest model, SVM models employing radial basis function (RBF) and polynomial kernels, and the MLP neural network exhibited relatively lower classification performance. The superior performance of the CART algorithm may be attributed to its simplicity, robustness, and ability to handle nonlinear relationships efficiently, particularly when working with relatively small datasets. Furthermore, CART provides highly interpretable decision rules, making it suitable for environmental applications where model transparency is desirable. The final soil salinity map indicated that approximately 13.5% of the rangelands in Ramhormoz County are affected by high salinity levels. Areas with elevated salinity were primarily concentrated in the northern, southern, and western parts of the study area. These spatial patterns were consistent with field observations, supporting the reliability of the generated salinity map and the adopted methodological framework.
Conclusion The results of this study demonstrate that integrating Landsat 9 remote sensing data with machine learning algorithms can provide an effective approach for soil salinity assessment and mapping in arid and semi-arid rangelands. Among the evaluated algorithms, the CART model exhibited the best overall performance and proved to be a reliable tool for soil salinity classification under the environmental conditions of the study area. The findings highlight the importance of selecting appropriate machine learning models when developing soil salinity maps and suggest that decision tree-based approaches may outperform more complex algorithms when sample sizes are limited. Future research should investigate the integration of higher spatial resolution optical and radar imagery, larger field datasets, and advanced deep learning techniques to further improve the accuracy and robustness of soil salinity mapping in rangeland ecosystems.

Abstract Abstract
Introduction: Cadmium is considered a critical heavy metal due to its high mobility in the environment, which poses serious risks to ecosystem health through its entry into the food chain. Accordingly, it is necessary to achieve efficient methods for stabilizing this element in contaminated soils. The present study aimed to synthesize and compare the performance of natural zeolite nanoparticles and microparticles and biochar in stabilizing cadmium in soil.

Materials and Methods: For this purpose, the incubation experiment was carried out in a completely randomized factorial design with three replications and a time period of two months. The structural characteristics of the particles produced through the planetary ball mill were determined using XRD, BET, SEM and FTIR analyses. The effect of different treatments (including zeolite and biochar in both nano and micro scales with levels of 1 and 3 wt%) on the cadmium adsorption and changes in its chemical forms was evaluated.

Results and Discussion: The findings indicated that all modifiers improved the cadmium fixation efficiency with increasing dosage. Also, particle size had a significant effect on the adsorption capacity; so that the performance of zeolite nanoparticles and biochar was superior to the micro versions. Among them, the 3% nanobiochar treatment showed the highest fixation potential with an efficiency of 95.27%. The study of cadmium chemical species showed that due to the high pH and the poverty of organic matter in the soil, the abundance of soluble, exchangeable and organic forms was lower and the carbonate form was the dominant species due to the high presence of calcium carbonate in the soil. Finally, the results confirmed that the use of modifiers, especially nanobiochar, in addition to reducing the availability of cadmium, leads to an effective redistribution of the chemical species of this element in the soil.

Conclusion: The finding of this study demonstrates that producing nano-sized zeolite and biochar via ball milling is an efficient and cost -effective strategy for enhancing cadmium stabilization cadmium in contaminated soils. FTIR analysis confirmed that mechanical size reduction of biochar does not alert the type of functional groups present; however, the number of active functional groups increase as particle size decreases. XRD results showed that mechanical milling reduces the crystallinity of natural zeolite, which contributes to improved reactivity. Nanozeolite through higher cation exchange capacity and specific surface area and nano biochar due to the high amount of functional groups, higher cation exchange capacity, and specific surface area reduces cadmium in the soil compared to micro particulate. Among the modifiers, Nano Biochar performed better. Also, each of the treatments had different effects on the distribution of cadmium chemical forms in the soil. Nano biochar at 3% application level, with the greatest reduction in exchangeable forms and carbonate and increasing the organic and residual fraction, has a greater effect on the redistribution of cadmium in the soil and by converting cadmium into inaccessible forms, partially reduces the concentration of mobile cadmium and reduces its availability. It can be concluded that the use of nano-biochars of corn produced at 400 °C at higher doses is useful for stabilizing cadmium in calcareous contaminated soils.

Abstract Introduction Microplastics are emerging soil contaminants that can alter soil physicochemical properties and consequently influence the mobility, bioavailability, and chemical speciation of heavy metals. Because the environmental behavior of metals is governed primarily by their chemical speciation rather than their total concentrations, understanding the factors controlling metal speciation is essential for predicting their environmental fate and ecological risk. Among these factors, soil pH and dissolved organic carbon (DOC) are the primary regulators of metal speciation, and both microplastics and plant growth-promoting rhizobacteria (PGPR) can modify these properties, thereby influencing the geochemical behavior of metals. Although increasing attention has been paid to the interactions between microplastics and heavy metals and to the role of PGPR in regulating metal behavior in the rhizosphere, most previous studies have focused on the adsorption of metals onto microplastic surfaces, whereas their effects on metal speciation in soil solution remain poorly understood. Therefore, the present study investigated the interactive effects of low-density polyethylene (LDPE), polylactic acid (PLA), and two PGPR strains on the chemical speciation of nickel and lead in the soil solution of a contaminated calcareous soil.
Materials and Methods A greenhouse experiment was conducted as a factorial arrangement in a completely randomized design with three replications. Artificially contaminated calcareous soil containing Pb, Ni, Cd, and Zn was amended with two types of microplastics, low-density polyethylene (LDPE) and polylactic acid (PLA), each applied at two rates (1 and 5%, w/w), and inoculated with two plant growth-promoting rhizobacterial strains, Pseudomonas fluorescens and Azospirillum lipoferum. Maize (Zea mays L.) seeds were sown in the treated soils and grown for 45 days. After harvest, soil pH, electrical conductivity (EC), dissolved organic carbon (DOC), dissolved Pb and Ni concentrations, and the ionic composition of the soil solution were determined.
The chemical speciation of Pb and Ni was predicted using the geochemical equilibrium model Visual MINTEQ based on soil solution chemistry, and the resulting data were statistically analyzed using SAS software.
Results and Discussion The results demonstrated that microplastics significantly altered the chemical speciation of nickel and lead through changes in soil chemical properties, particularly pH and DOC. LDPE decreased soil pH, shifting the speciation equilibrium toward higher proportions of free Ni²⁺ and Pb²⁺ ions and lower proportions of carbonate and hydroxide species. Under the 5% LDPE treatment, the proportion of Ni²⁺ was approximately 53–62% of total dissolved Ni, whereas Pb²⁺ accounted for approximately 11–21% of total dissolved Pb. In contrast, PLA increased both pH and DOC, promoting the formation of carbonate, hydroxide, and dissolved organic matter complexes (NiDOM and PbDOM). These findings are consistent with the well-established role of pH in regulating metal speciation and the contribution of DOC to metal–organic complexation. Furthermore, inoculation with PGPR, particularly Azospirillum lipoferum, further enhanced the formation of organic complexes under PLA treatments by increasing DOC. Analysis of the relative distribution of aqueous species revealed that Ni²⁺ (26–62%) and PbCO₃ (34–57%) were the predominant species of nickel and lead, respectively. Despite the enhanced formation of organic complexes under PLA treatments, nickel remained predominantly in its free ionic form, whereas lead exhibited a much stronger tendency to form organic complexes. In the PLA treatments inoculated with PGPR, the proportion of Ni²⁺ (34.76%) remained higher than that of NiDOM (21.97%), whereas for lead, the proportion of PbDOM (24.29%) was substantially higher than that of Pb²⁺ (1.23%). Overall, LDPE increased the proportion of free ionic species of both metals, suggesting a potential increase in their bioavailability, whereas PLA altered their geochemical behavior primarily by promoting complexation with dissolved organic matter, particularly for Pb. Conclusion The findings of this study demonstrated that microplastic type significantly altered the chemical speciation of nickel and lead in calcareous soil through changes in soil pH and DOC. By lowering soil pH, LDPE increased the proportion of free metal ions, thereby potentially enhancing their bioavailability. In contrast, PLA promoted the formation of metal–organic complexes, particularly for Pb, through increases in both pH and DOC. Furthermore, PGPR amplified the effect of PLA by stimulating DOC accumulation, resulting in greater metal–organic complexation. Overall, these findings highlight that both microplastic type and microbial activity are key drivers of the geochemical behavior and environmental fate of nickel and lead in the soil solution phase.

Abstract Introduction: Evaporation from open water surfaces is a primary cause of water resource depletion in arid and semi-arid regions, sometimes resulting in the loss of over 50 percent of annual water reserves. Iran, characterized predominantly by an arid to semi-arid climate, is significantly affected by this phenomenon; reports indicate that between 2003 and 2019, the country lost approximately 211 cubic kilometers of its total water reserves—an amount exceeding twice the nation's annual water consumption. Agricultural ponds and reservoirs are particularly vulnerable to such losses due to constant exposure to direct sunlight; indeed, annual evaporation from small agricultural ponds in the Middle East is estimated to exceed 2,400 million cubic meters. Exacerbating this challenge is the fact that nearly 47.8 percent of the country's irrigation pumps still operate on diesel engines, which—in addition to high operating costs—result in significant greenhouse gas emissions. Floating photovoltaic (FPV) systems have emerged as a dual-purpose technology capable of simultaneously reducing surface evaporation and generating clean electricity to power well motors and irrigation pumps. Although previous studies have reported significant evaporation reduction through the use of floating photovoltaic covers—ranging from approximately 40% to over 90%, depending on the coverage extent and layout—the precise impact of panel installation positioning (submerged, flush with the water surface, or elevated above the water) on evaporation reduction in arid climates has not been systematically investigated. In addition to evaporation reduction, the electrical efficiency of the panels under each installation configuration was evaluated as a complementary finding. The present study was designed and conducted to address this knowledge gap.

Materials and Methods This experimental study was conducted during the summer of 2024 (August 6 to September 5) at the Faculty of Agriculture, Ferdowsi University of Mashhad, located in the arid climate of northeastern Iran. Four identical tanks (65 × 79 × 30 cm) were each placed inside a 1 × 1 × 1 m polystyrene enclosure to eliminate the effect of lateral solar radiation on the tank walls. Four treatments were defined: a control tank without a panel; a panel submerged 2 cm below the water surface; a panel flush with the water surface; and a panel positioned 8 cm above the water surface. Four identical 20-watt polycrystalline panels (model Zytech Solar ZT-20P, labeled A through D) were installed horizontally, facing south with a zero-degree tilt angle. To account for panel-specific effects, the positions of the panels were rotated every three days between treatments, resulting in six independent replicates for each treatment. Each panel was connected to an identical resistive circuit—consisting of a 20-watt incandescent bulb—to simulate actual operating temperature conditions. The voltage and current of each panel were measured individually and in immediate succession using a Victor VC97 digital multimeter, with the panel connected to an identical 20-watt incandescent lamp (acting as a resistive load). Voltage was measured across the load, while current was measured in series within the circuit. Besides, electrical output power was calculated . Daily evaporation rates were measured using a digital scale (model FX-3000iWP, precision ±0.01 g) and a 1000 mL graduated beaker, while environmental parameters—including air temperature, relative humidity, and solar radiation intensity—were simultaneously recorded using a Lutron HT-3027SD data logger and the PVGIS database. Assumptions regarding data normality (Shapiro-Wilk test) and sphericity (Mauchly's test) were assessed, and the data were analyzed using repeated measures ANOVA (RM-ANOVA) in SPSS software (version 27), accompanied by pairwise comparisons with Bonferroni correction.
Results and Discussion The average daily evaporation in the control treatment was 4.39 mm, representing the highest value recorded among all treatments. The submerged panel treatment resulted in only a limited reduction (15.7%), bringing the evaporation rate to 3.70 mm; this was likely due to heat transfer from the panel to the water body, which partially offset the shading effect. The treatment with the panel positioned above the water surface showed a reduction of 45.63% (2.38 mm), whereas the treatment with the panel flush with the water surface yielded the greatest reduction (52.37%, or 2.09 mm) as well as the most consistent performance throughout the monitoring period. Repeated measures ANOVA (RM-ANOVA) confirmed statistically significant differences among the treatments (F = 159.17, df = 3, p < 0.001), and Bonferroni post-hoc tests revealed that all panel treatments differed significantly from the control (p < 0.001). The numerical difference between the water-level treatment and the above-water treatment was negligible (0.296 mm) and statistically non-significant (p = 0.065); however, the water-level treatment exhibited greater daily stability throughout the experiment. In contrast to the study by Lopez et al., which achieved a 37% reduction in evaporation using 50% surface coverage on water reservoirs in semi-arid Brazil, the present study recorded a 52.37% reduction with only 42% coverage—a result that exceeds predictions based on a linear relationship between coverage extent and evaporation, highlighting the importance of optimal installation positioning regardless of the surface coverage area. Installation position also significantly affected the panel's electrical efficiency (p < 0.01): the water-surface-level treatment achieved the highest mean efficiency (10.9%), followed by the above-water (10.7%), control/ground-based (10.5%), and submerged (10.2%) treatments; despite recording the highest open-circuit voltage, the submerged treatment exhibited lower efficiency due to optical losses caused by light absorption by the water. The alignment of enhanced evaporation reduction and optimal electrical performance in the water-surface-level treatment confirms the potential of this configuration for dual-use applications.
Conclusion These findings indicate that the installation configuration of floating photovoltaic panels is a decisive factor in their performance—regarding both evaporation reduction and electricity output—under arid climatic conditions. A water-surface-level configuration maximized evaporation reduction while simultaneously yielding the highest electrical efficiency; consequently, this approach is recommended as a practical, dual-purpose solution for agricultural water reservoirs in arid regions of Iran (such as Khorasan Razavi Province), as it enables water resource conservation alongside energy generation for irrigation pumping. Future research should focus on evaluating the structural durability of these systems at an operational scale, investigating their long-term impacts on water quality and reservoir ecosystems, and conducting techno-economic feasibility analyses of the technology.

Abstract Introduction: Strawberry (Fragaria × ananassa Duch.) is one of the most valuable and widely cultivated berry crops in the world due to its high nutritional value, appealing flavor, and economic importance. In hydroponic systems, where nutrient solution is the sole source of mineral supply, precise management of micronutrients is essential for optimal plant growth and fruit quality. Among these micronutrients, copper (Cu) and zinc (Zn) play key roles in photosynthesis, enzyme activation, protein synthesis, and antioxidant defense. However, conventional mineral salts often show limited availability and may cause imbalances, whereas Amino-chelated (AC) forms can improve nutrient uptake and reduce toxicity. Therefore, this study aims to evaluate the combined foliar application of copper and zinc amino-chelates on the vegetative and reproductive traits of two strawberry cultivars, Sabrina and Camarosa, under hydroponic conditions.
Materials and Methods: To evaluate and compare the effects of Zn and Cu amino-chelates on the morphological characteristics of strawberry, a factorial experiment based on a completely randomized design with three replications was conducted in the greenhouse of the Faculty of Agriculture at Shahid Chamran University of Ahvaz. In January 2025, strawberry runners of ‘Camarosa’ and ‘Sabrina’ cultivars were obtained from a nursery in Kordestan province and transferred to the greenhouse. The plants were disinfected with Mancozeb fungicide and planted in growing beds (130 cm long, 25 cm wide, and 10 cm deep) containing a 70:30 mixture of coco-peat and perlite. The Rash nutrient solution was used for plant nutrition, and treatments commenced one month after plant establishment. The experimental treatments included foliar application of Cu AC at two levels (0 and 4 ml/L) and Zn AC at three levels (0, 2, and 4 ml/L) on both ‘Sabrina’ and ‘Camarosa’ cultivars. Foliar spraying was repeated monthly. Measured parameters included vegetative and reproductive characteristics, as well as photosynthetic pigments.
Results and Discussion: The study revealed that foliar application of Zn and Cu amino-chelates significantly improved vegetative growth, reproductive yield, and photosynthetic pigments in both 'Sabrina' and 'Camarosa' strawberry cultivars. Optimal responses were generally observed at 2-4 mL/L concentrations, highlighting enhanced biomass, fruit quality, and gas exchange efficiency. The highest shoot fresh weight (26 g) and dry weight (6.36 g), as well as the highest root fresh weight (16.33 g) and dry weight (3.79 g), were obtained in the Sabrina cultivar under the treatment of 2 mL L⁻¹ Zn amino chelate without the application of Cu amino chelate. The treatment consisting of 4 mL/L Cu amino-chelate combined with 2 mL/L Zn amino-chelate resulted in the highest levels of chlorophyll b (0.240 mg/gfw) and photosynthetic rate (8.57 µmol/m²/s). Zn and Cu amino-chelates improved vegetative growth, reproductive performance, and physiological traits in strawberry by enhancing nutrient uptake, chlorophyll synthesis, gas exchange, and growth-related metabolic processes. Zn likely promoted growth through its role in auxin biosynthesis, enzyme activation, and reproductive development, while Cu contributed to structural stability and root dry matter through lignin synthesis.
Conclusion: Overall, the results indicated that foliar application of Zn and Cu amino-chelates positively affected the vegetative, reproductive, and photosynthetic pigment indices in both ‘Sabrina’ and ‘Camarosa’ cultivars, with the treatments showing a greater increase in morphological traits for ‘Sabrina’ compared to ‘Camarosa

Abstract Introduction Jujube (Ziziphus jujuba Miller) is an economically and nutritionally valuable fruit crop, particularly in arid and semi-arid regions. Drying is a critical postharvest operation that extends shelf life, facilitates storage and transportation, and increases the commercial value of jujube. However, precise determination of the drying status is essential for achieving consistent product quality. Inadequate drying may result in excessive residual moisture and reduced storage stability, whereas over-drying can cause undesirable changes in product quality and unnecessary energy consumption. Conventional approaches for determining drying status, including moisture measurement, weight loss, and visual inspection, are often time-consuming, destructive, or dependent on operator expertise. Computer vision combined with deep learning offers a rapid and non-destructive alternative for automated assessment of agricultural products. Although previous studies have demonstrated the effectiveness of deep learning for jujube maturity grading, defect detection, quality classification, and non-destructive evaluation, relatively limited attention has been given to the direct classification of jujube drying statuses using RGB images. Therefore, this study aimed to develop a dedicated deep convolutional neural network, termed DCNN_Jujube, for automatic classification of different drying statuses of Ghaenat jujube kernels and to compare its performance with transfer-learning and conventional machine-learning approaches. In this study, drying was performed using a hot-air dryer at 70 °C with an air velocity of 1 m/s.
Materials and Methods A dataset comprising 1,840 images of Ghaenat jujube kernels was established and categorized into four drying-status classes: fresh, semi-dried, dried, and excessively dried. The dataset consisted of 1,280 training images and 560 validation images, while an independent test set of 160 images was used for final performance evaluation. All images were cropped and resized to 512 × 512 pixels. To increase training-data variability and mitigate overfitting, augmentation techniques, including flipping, cropping, translation, scaling, and rotation, were applied. The proposed DCNN_Jujube architecture was specifically designed for the target four-class classification problem. The network comprised ten convolutional blocks and two fully connected blocks and incorporated batch normalization, ReLU activation, max-pooling, and dropout layers. A Softmax layer was used for final classification. For comparative evaluation, two pretrained architectures, InceptionV3 and InceptionResNetV2, were fine-tuned using the jujube dataset. In addition, conventional machine-learning methods based on handcrafted HOG and SIFT features were evaluated using support vector machine (SVM), k-nearest neighbors (KNN), and logistic regression classifiers. Model performance was evaluated using accuracy, precision, recall, F1-score, confusion matrix, and receiver operating characteristic (ROC) analysis.
Results and Discussion The proposed DCNN_Jujube achieved the highest classification performance, with an overall test accuracy of 97.00%. The corresponding precision, recall, and F1-score were 96.91%, 96.88%, and 96.88%, respectively. These results demonstrate the strong capability of the proposed architecture to discriminate among the four drying statuses. The fine-tuned transfer-learning models also achieved high classification performance. InceptionV3 and InceptionResNetV2 obtained test accuracies of 95.63% and 96.25%, respectively. Although both pretrained models provided competitive results, their performance was slightly lower than that of DCNN_Jujube. This finding suggests that a task-specific architecture can learn visual representations more closely associated with the characteristics of jujube drying than general-purpose pretrained models. In contrast, the conventional machine-learning approaches based on HOG and SIFT features achieved lower classification accuracies, ranging approximately from 85.00% to 92.50%. The lower performance of these methods may be related to the limitations of handcrafted features in representing the complex and progressive changes in color, texture, shape, shrinkage, and surface appearance that occur during drying. In comparison, the proposed convolutional network automatically learned hierarchical and discriminative features directly from the images. The confusion matrix demonstrated that the majority of samples were correctly assigned to their corresponding drying statuses. Some misclassification between adjacent statuses was observed, which is reasonable because drying is a continuous process and the visual differences between consecutive statuses may be subtle. The ROC analysis further confirmed the strong discriminative capability of DCNN_Jujube. Overall, the findings demonstrate that RGB image-based deep learning can provide an effective and relatively simple approach for non-destructive classification of jujube drying statuses.
Conclusion A dedicated deep convolutional neural network, DCNN_Jujube, was successfully developed for the automatic classification of four drying statuses of Ghaenat jujube kernels using RGB images. The proposed model achieved a test accuracy of 97.00%, with precision, recall, and F1-score of 96.91%, 96.88%, and 96.88%, respectively, and outperformed both fine-tuned transfer-learning models and conventional machine-learning approaches based on handcrafted features. The results demonstrate that a task-specific deep-learning architecture can effectively extract discriminative visual characteristics associated with progressive jujube drying and provide accurate multi-class classification. The proposed RGB-based approach has potential as a rapid, non-destructive, and relatively cost-effective tool for drying-status assessment, online quality monitoring, and intelligent process control. Further research should evaluate the model using larger and more diverse datasets covering different cultivars, drying conditions, illumination environments, and imaging configurations, followed by validation under real industrial processing conditions.