1. Introduction
Precision Agriculture
Precision agriculture (PA), also known as site-specific management, offers a sustainable solution to the growing demand for food. By integrating agricultural sciences with information technology, this approach analyzes spatial and temporal variations within fields [1]. PA techniques have been developed to conserve natural resources such as water and soil, and optimize the use of fertilizers, irrigation, pesticides, and weed identification, among others, while reducing production costs [2,3,4] The modernization of agriculture, driven by embedded electronics and remote sensing, has revolutionized traditional methods and fostered industry growth [1], adapting agricultural practices to the specific needs of different field areas.
The contribution of digital agriculture to precision agriculture lies in the application of digital technologies to collect, interpret, and communicate data that support decision-making on farms and throughout the production chain, characterizing a data-driven agricultural approach. In this context, precision agriculture has advanced through the use of smart machinery, sensors, drones, and satellites, which generate high-resolution data analyzed by machine learning algorithms. This analysis enables the accurate detection of plant diseases and nutrient deficiencies. Practical examples, such as automated cattle feeding based on weight, highlight the importance of data collection and processing. Thus, digital agriculture enhances the potential of precision agriculture by promoting greater efficiency and improving decision-making in the field.
Remote sensing plays a crucial role in precision agriculture, serving as a tool for acquiring data and information over large cultivated areas in a short period. This technology enables the analysis of target attributes based on their interaction with the electromagnetic spectrum [5]. Remote sensors are equipment designed to measure a wide range of data, including the reflectance of the electromagnetic spectrum coming from an object or plant.
There are two main types of hyperspectral sensors: imaging sensors and those that provide spectral values for each wavelength. With decreasing equipment costs, imaging-based sensors have become increasingly utilized, particularly when mounted on unmanned aerial vehicles (UAVs), expanding their application in aerial mapping and precision agriculture. Hyperspectral sensors capture a broad range of wavelengths, enabling high spatial (visual detail) and spectral (wavelength differentiation) resolution. This capability allows for the precise identification and monitoring of specific plant, soil, and surface characteristics [6], such as crop physiological monitoring, water stress detection, and nutrient assessment.
UAVs equipped with high-resolution imaging sensors are particularly effective in data collection, surpassing satellite imagery in detecting crop stress and analyzing soil attributes [7]. UAVs provide higher spatial resolution images and are not affected by cloud cover, ensuring continuous operation. These data are essential for developing decision-support systems, facilitating intelligent pest control, fertilization, and irrigation management [8].
Plant leaves under stress reflect specific conditions, making them efficient for spectral characterization, as they contain substances such as chlorophyll, which indicate the plant’s phenological status [5]. The visible and infrared regions of the electromagnetic spectrum exhibit characteristic reflectance curves [9]. Hyperspectral sensors allow for continuous spectral sampling, covering wavelengths mostly ranging from 350 to 1100 nm, providing greater efficiency compared to multispectral sensors, which only capture electromagnetic reflectance in specific bands [5,10,11].
Chlorophyll primarily absorbs light in the red (600–700 nm) and blue (400–500 nm) spectral bands while reflecting green light (500–600 nm). Additionally, carotenoid pigments reflect more light in the yellow-orange region (500–600 nm), whereas water molecules influence reflectance in the near-infrared region (700–1300 nm). When plants experience stress, such as water or nutrient deficiency, significant changes occur in reflectance, particularly in the near-infrared and red regions [12].
In a recent study by [11], spectral attenuations were observed at specific wavelengths, particularly in the green spectrum (around 550 nm), forming a cluster between 530 and 560 nm, and in the red-edge spectrum (around 750 nm), forming another cluster between 690 and 820 nm. These results indicate that healthy leaves exhibit a typical pattern of light absorption and reflection, with chlorophyll playing a crucial role in absorbing red and blue wavelengths while reflecting green and near-infrared light [12].
In the study by [13], the spectral bands most sensitive to leaf nitrogen content in degraded temperate vegetation were identified in the red and near-infrared (NIR) regions, specifically at wavelengths of 468, 623, 624, 633, 652, 657, 668, 818, 821, 842, 937, and 938 nm. The nitrogen-sensitive bands were primarily found in the green, green-yellow edge, mid-red, and NIR regions of the spectrum. Similarly, [11,13] also suggested that the identification of nitrogen-sensitive wavelengths should be conducted across the entire spectral range rather than being limited to specific bands. This approach would enhance the accuracy of nitrogen content estimation in leaves, as it captures a broader range of spectral variations, leading to greater precision in detecting nutrient deficiencies.
The relationship between phenological stage, chlorophyll content, and spectral readings is well illustrated during the vegetative growth of plants. At this stage, chlorophyll production is high, resulting in greater light absorption in the red region (600–700 nm) and, consequently, lower reflectance. As the plant progresses to senescence, chlorophyll levels decline, leading to increased reflectance in the same spectral range. This phenomenon can be detected by spectral sensors, which are capable of monitoring these changes non-destructively [14].
The fundamental premise is to move beyond managing fields based on average conditions and instead consider the spatial variability of factors influencing crop productivity. Precision agriculture techniques must necessarily incorporate the concept of “spatial variability” [15]. Therefore, this review aims to discuss research, terminology, and challenges in digital agriculture, with the goal of optimizing plant nutrition effectively, precisely, at a low cost for farmers, and contributing to sustainable agriculture.
In this article, we first address the contribution of artificial intelligence to precision agriculture (Section 2), exploring its role in advancing modern farming practices. Next, we discuss digital agriculture and precision agriculture, with a focus on the related concepts, differences, and contributions to plant nutritional assessment (Section 3), which is subdivided into three parts: Section 3.1, ‘What Is Digital Agriculture?’; Section 3.2, ‘How Does It Differ from Precision Agriculture?’; and Section 3.3, ‘Why Can Digital Agriculture Make Significant Contributions to Assessing Plant Nutritional Status?’.
In Section 4, we present key machine learning algorithm models applied to plant nutritional diagnostics in agricultural systems. Following this, Section 5 discusses the main challenges of digital agriculture, highlighting current limitations and opportunities for further development. Section 6, entitled ‘From Leaf to Harvest: How Interactions Between Spectral Readings, Leaf Nutrient Content, and Yield Affect the Robustness of Predictive Models’, explores the importance of validating plant nutritional diagnostics through crop productivity analysis. Finally, we present our final considerations (Section 7), where the main points discussed throughout the article are summarized, and future research directions are suggested. The references used throughout the article are listed at the end.
2. Contribution of Artificial Intelligence to Precision Agriculture
The integration of artificial intelligence (AI) into precision agriculture enhances farming practices by applying data-driven insights for more effective decision-making [16]. AI technologies, such as machine learning and computer vision, process vast amounts of data from various sources, including sensors, UAVs, and satellite imagery [17]. This enables detailed monitoring of crop health, soil conditions, and multiple environmental factors [1].
In recent years, artificial intelligence has gained prominence, largely due to the evolution of data formats, which have shifted from traditional structured formats to unstructured, semi-structured, and heterogeneous architectures, each with distinct characteristics [17]. Advances in large language models (LLMs), machine learning (ML), and neural networks (NNs) have continued to progress [18].
Artificial intelligence does not replicate human cognitive processes but rather simulates aspects of thinking and decision-making through data processing and algorithms [19,20]. Human intelligence is defined as the ability to think, plan, and solve complex problems [21]. Similarly to human cognition, AI has multiple interpretations and is often defined as “the ability of machines to perform tasks that typically require human intelligence” [19].
Remote sensing, combined with artificial intelligence, enables real-time data collection and analysis, optimizing and improving decision-making processes [22,23]. “Smart agriculture” emphasizes the application of AI and data science [24]. Machine learning (ML) algorithms facilitate the processing of sensor-derived data, enabling fast and automated analyses [25]. These algorithms can handle datasets of varying sizes and complexities, depending on model efficiency [26,27].
Artificial intelligence (AI) plays a fundamental role in precision agriculture by simulating human intelligence through machine learning. However, AI does not replace human intelligence, as proper management is required. Moreover, the interaction between artificial intelligence and human expertise is essential for optimizing results. While AI can rapidly process and analyze large volumes of data, the experience and judgment of agricultural professionals and farmers remain indispensable for making informed decisions based on the generated analyses. This collaboration between machines and humans enhances agricultural practices, enabling a more precise and effective approach to crop management.
3. Digital Agriculture and Precision Agriculture: Concepts, Differences, and Contributions to Plant Nutritional Assessment
3.1. What Is Digital Agriculture?
The global transition to the digital era is crucial for society, as most daily activities are driven by digital technologies aimed at increasing productivity and efficiency in food, fiber, and bioenergy production [28,29,30]. However, the unequal access to digital techniques and equipment in both rural and urban areas represents a significant barrier, threatening technological efforts to combat hunger, ensure sustainable agriculture, and promote innovations aligned with the Sustainable Development Goals (ODS, 02 e 09) [31].
Remote regions, particularly in rural areas of less developed countries, are often disconnected from technological advancements. Digital transformation in these areas is essential for shifting from subsistence farming to production integrated into the global market [28]. By adopting digital technologies, small-scale farmers can efficiently manage agricultural systems [32]. The application of inputs in the field becomes more precise, utilizing collected data to determine the appropriate amount, timing, and location for nutrient, water, and agrochemical applications, contributing to sustainable and efficient management [33,34].
One of the most common devices in modern agriculture is the mobile phone, which, through applications, facilitates information sharing. However, it is important to emphasize that connectivity is not solely dependent on mobile devices but rather on the availability of mobile network signals. These signals enable communication and telemetry, which are fundamental to digital agriculture. Furthermore, agricultural machinery has already been optimized with specialized software [35,36], allowing for even greater integration between technology and farming practices.
The concept of digital agriculture integrates various technologies, such as automation, robotics, the Internet of Things (IoT), remote sensing, big data, artificial intelligence, and data analytics into agricultural practices, ensuring sustainability [8,37,38]. The authors of [39] pointed out, to avoid confusion, that in this context, digital agriculture refers to the implementation of digital technologies in cropping and livestock systems to collect, interpret, and communicate data that support decision-making on farms and throughout the supply chain. In summary, it is a data-driven approach to agriculture [40].
This approach aims to maximize productivity through integrated sensors in UAVs, connected to mobile devices and agricultural machinery, enabling real-time monitoring of livestock, soil, water, and crops, as well as human interactions [30,41]. A notable example of digital agriculture is cattle monitoring in pastures: as an animal passes through a gate, it is identified and weighed, prompting an automated system to dispense a precise amount of feed. Another example is agricultural fleet management, where a tractor’s engine temperature can be remotely monitored from a control center, independent of maps or production variables. The collection and interpretation of these data enable informed decision-making, assisting both in historical analysis and future predictions.
3.2. How Does It Differ from Precision Agriculture?
In digital agriculture, precision agriculture plays a key role, particularly in addressing spatial variability, such as fluctuations in soil nutrient levels across different locations. Additionally, precision agriculture accounts for temporal variability, referring to changes in agricultural conditions over time. These include climate fluctuations, crop growth variations, yield productivity shifts, and resource availability changes, such as water and nutrients.
Precision agriculture relies on advanced equipment, including machinery, sensors, UAVs, and satellites, to collect high-resolution data. Artificial intelligence (AI) processes this data through machine learning algorithms, generating predictive models. Once trained on these datasets, agricultural machinery can accurately detect plant diseases and nutrient deficiencies in the field. The earlier example of automated cattle feeding based on weight highlights how data collection and processing are fundamental in this context. Thus, digital agriculture integrates precision agriculture technologies with AI-driven analytics, optimizing agricultural productivity and providing farmers with a comprehensive overview of crop and farm conditions, thereby supporting informed decision-making.
3.3. Why Can Digital Agriculture Make Significant Contributions to Assessing Plant Nutritional Status?
In recent years, efforts to expand digitalization in agriculture—such as nutrient status sensors (whether mounted on UAVs or stationary) combined with predictive modeling—have yielded positive results in rural areas [42]. Notably, these advancements have been applied to plant nutritional assessment, where hyperspectral electromagnetic spectra have been utilized to predict nitrogen, chlorophyll, and carotenoid content (Table 1). Among predictive models, Random Forest (RF) has emerged as the most effective, owing to its ability to process large datasets and detect complex spectral patterns, leading to highly accurate nutrient estimations [11].
On the other hand, to estimate nitrogen (N), phosphorus (P), and potassium (K) levels, he study by the authors in [43] used predefined spectral band imagery to calculate various vegetation indices, including NDVI (Normalized Difference Vegetation Index), NDRE (Normalized Difference Red Edge Index), GNDVI (Green Normalized Difference Vegetation Index), SAVI (Soil-Adjusted Vegetation Index), MSAVI (Modified Soil-Adjusted Vegetation Index), MCARI (Modified Chlorophyll Absorption in Reflectance Index), EVI (Enhanced Vegetation Index), and SCCCI (Simplified Canopy Chlorophyll Content Index) (Table 1).
In the study mentioned above, in addition to testing different algorithms, three distinct data input methods (spectral bands, vegetation indices, and their combinations) were analyzed to determine which type of spectral data would yield the highest classification accuracy (Table 1). The input configuration that showed the best performance for machine learning algorithms was the use of spectral bands, with the J48 (decision tree) and SVM (support vector machine) algorithms standing out in predicting N, P, and K levels [43], confirming the findings of [11] mentioned in this review (Table 1).
Another significant study investigates the prediction of boron deficiency or toxicity in eucalyptus. Several vegetation indices were analyzed, including CCCI (Canopy Chlorophyll Content Index), CIgreen (Green Chlorophyll Index), EVI (Enhanced Vegetation Index), HMSSSI (Heavy Metal Stress Sensitive Index), MTCI (MERIS Terrestrial Chlorophyll Index), NDRE (Normalized Difference Red Edge Index), and NDVI (Normalized Difference Vegetation Index). Among these, NDVI proved to be the most effective in distinguishing adequate boron levels in Eucalyptus spp., both in young and mature plants. Additionally, the CCCI and PRSI (Photochemical Reflectance Stress Index) indices were used to identify deficiencies [44].
Spectral band analysis showed that the red region contributes the most to characterizing boron-deficient levels, with red edge bands at 740 nm, 782 nm, and 864 nm, along with NIR at 835 nm and the water vapor band at 945 nm, which are essential for discriminating adequate boron levels [44].
Some nutritional states were identified through analyses of different parameters, such as wavelength evaluation and algorithm application [45] (Table 1), using predefined wavelengths: blue (465–485 nm), green (550–570 nm), red (663–673 nm), red edge (712–722 nm), and near-infrared (820–1000 nm). For this analysis, multiple regression algorithms were applied, including ElasticNet, Lasso regression, Linear SVM, PLSR, Random Forest, and Ridge regression. The gradient boosting regression algorithm proved particularly effective for small datasets, outperforming the other tested methods [45]. However, one of the main challenges of this research was determining the best model and wavelength for each evaluated nutrient, a question that remains unresolved in the study.
Reflected radiation provides valuable information about plant nutrition, and vegetation indices are used to quantify this reflectance. Examples include the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Red Edge Index (NDRE), as utilized by [46]. Studies show that high nitrogen doses (180 kg ha−1) are positively related to these vegetation indices, resulting in a better spectral response in maize. The equation presented in Table 1 represents the observation in the k-th block, assessed in the i-th variety, j-th nitrogen fertilization, and l-th harvest; μ is the overall mean; Bk is the block effect, considered fixed; Cl is the harvest effect, considered random; Vi is the variety effect, considered fixed; Nj is the nitrogen level effect, considered fixed; CVil is the harvest by variety interaction, considered random; CNjl is the harvest by nitrogen level interaction, considered random; VNij is the variety by nitrogen level interaction, considered fixed; CVNijl is the interaction between planting season, variety, and nitrogen, considered random; and ϵijkl is the random error associated with observation Yijk [46].
An interesting study by [47] used a Vis-NIR hyperspectral sensor operating in the 400 to 1000 nm spectral range in a laboratory setting to determine the levels of N, P, K, and Ca. The analysis was carried out using the partial least squares regression (PLSR) model, which can process large datasets generated by hyperspectral images. Wavelengths in the 400 to 450 nm, 530 to 540 nm, 630 to 780 nm, and 990 to 1000 nm ranges were crucial for predicting foliar nitrogen levels. The model also predicted foliar phosphorus levels in the 700 to 1000 nm range. For potassium prediction, the 730 to 1000 nm range presented a higher number of informative wavelengths compared to other nutrients. Meanwhile, wavelengths from 470 to 520 nm, 570 to 650 nm, 710 to 730 nm, 770 to 820 nm, and 860 to 1000 nm were used to estimate calcium levels [47].
To predict nitrogen levels, research was conducted using multispectral images [48,49,50]. The input data included spectral bands and vegetation indices, such as the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Red Edge Index (NDRE), the Green Normalized Difference Vegetation Index (GNDVI), and the Soil-Adjusted Vegetation Index (SAVI). These data were analyzed by various models, including REPTree (REPT), Random Forest (RF), K-nearest neighbor (K = 1, 5, and 10), SVM-RBF, Polynomial Support Vector Machine (SVMP), Linear Regression (LR), and RBF Regression. Among these, the Random Forest algorithm showed the best performance, with the vegetation indices contributing more to the model’s effectiveness than the spectral bands alone [49].
Among the 33 vegetation indices analyzed by [49], 12 exhibited regression coefficients greater than 0.5. These indices include the Canopy Chlorophyll Content Index (CCCI), Green Chlorophyll Index (CG), Red Edge Chlorophyll Index (CIrededge), Chlorophyll Vegetation Index (IVC), Green Soil-Adjusted Vegetation Index (GSAVI), Modified Soil-Adjusted Vegetation Index (MSAVI), Renormalized Difference Vegetation Index (RDVI), Soil-Adjusted Vegetation Index (SAVI), Simple Ratio 750/550 (SR750/550), Simple Ratio 800/550 (SR800/550), Triangular Vegetation Index (TriVI), and Wide Dynamic Range Vegetation Index (WDRVI). Only five of these indices showed correlation coefficients greater than 0.7 in linear and exponential regressions: MSAVI, SR750/550, SR800/550, TriVI, and WDRVI. The Random Forest (RF) model outperformed most other machine learning algorithms, achieving a coefficient of determination (R2) of 0.90, with the XGBoost model also providing comparable metrics [49]. This means that the model explains 90% of the variation in the predictive data, indicating high accuracy in the model’s predictions.
To investigate the relationship between nitrogen content in orange trees and surface reflectance obtained from imagery, the study in [48] applied spectral analysis algorithms to high-spatial-resolution images using multiple spectral bands. The analyzed spectral ranges included the green (510–590 nm), red (620–700 nm), red edge (725–745 nm), and near-infrared (750–830 nm) bands. Nitrogen content was classified into three categories: low (less than 27 g kg−1), medium (27 to 29 g kg−1), and high (more than 29 g kg−1). Each nitrogen class (low, medium, and high) had its spectral signature grouped accordingly.
These thresholds were determined based on research correlating foliar nitrogen content (g kg−1) with productivity increases (%) in field experiments [51]. A nitrogen content of 27 g kg−1 was associated with an approximate 100% increase in productivity, whereas above 29 g kg−1, the rate of increase began to decline. Nitrogen classification was performed using the Spectral Angle Mapper (SAM) algorithm, which used spectral curves as training data for the classifier. This algorithm outperformed others tested, achieving an overall accuracy of 87.5% and a kappa index of 0.75 [48].
4. Machine Learning Algorithm Models
In this section, we will discuss several machine learning algorithm models, including Artificial Neural Networks (ANNs), the M5P tree, REPTree, Random Forest, the Polynomial Kernel Support Vector Machine, and ZeroR. Artificial Neural Networks (ANNs) are computational models inspired by the structure and functioning of the human brain. They consist of interconnected units called neurons, organized into layers. These networks are widely applied in solving complex problems, including pattern recognition, data classification, and predictions across various fields, such as artificial intelligence, image processing, and data analysis [52,53].
The M5P tree is a more advanced model proposed by [54] to solve both linear and nonlinear regression problems [55]. This model uses unidimensional functions at the final nodes of the tree to perform statistical predictions, applying linear regression models at the terminal nodes [54,56].
REPTree (Reduced Error Pruning Tree) is a machine learning algorithm that constructs decision or regression trees based on entropy calculations, adjusting the model to minimize errors caused by data variance. It uses the reduce-error pruning process to prevent overfitting, enhancing the tree’s generalization capability [56].
Random Forest (RF) models are a powerful and innovative machine learning approach capable of mapping nonlinear relationships between input variables and their respective responses [57,58]. Among the most notable features of this technique are its relatively small number of advanced hyperparameters, automatic estimation of generalization errors, ability to handle missing values, and robustness against overfitting [57,59].
The Polynomial Kernel Support Vector Machine (SVMP) is an extension of SVM designed to solve complex problems by mapping data into higher dimensions using a polynomial kernel [60,61]. The polynomial kernel transforms non-linearly separable data into a higher-dimensional space, allowing SVM to identify more complex decision boundaries and better adapt to nonlinear patterns [60].
ZeroR is a machine learning model that serves as a basic benchmark for comparison with other models. It is a simple classifier that does not use any information about data features, relying solely on the most frequent class to make predictions [62]. ZeroR is a useful tool for initial data analysis, but its simplicity means it should be complemented with more sophisticated models to achieve more accurate and meaningful predictions.
From this perspective, these models have been applied in several agricultural projects seeking to solve complex tasks. There is a wide use of machine learning (ML) in agriculture with spectral data due to a combination of factors involving data complexity, prediction potential, and gains in efficiency and accuracy. For example, [63] proposed that using machine learning models with spectral data offers an efficient and low-cost long-term solution for determining caffeine in coffee beans, successfully achieving significant values using SVM combined with spectral information from ground and roasted beans to identify clones, with correlation coefficient values close to 0.7.
Another relevant application of the use of ML models was their use in the prediction of leaf nitrogen and the height of corn plants using multispectral data. The authors concluded that the use of RF is suitable for predicting both agronomic variables in corn, making it easier for farmers to monitor their plants, improving their production rates [64]. The use of a multispectral UAV sensor with a thermal sensor associated with the Random Forest algorithm proved to be efficient when applied to the RF model to predict corn productivity, providing the producer with important information for end-of-harvest policies with the crop still in the middle of the cycle [65]. The use of machine learning algorithms can accurately estimate the diameter (DBH) and height (H) of trees in mixed forest stands using hyperspectral data, especially when using DT, M5P, and SVM, facilitating forest inventories, for example, making the work more efficient [66].
In addition to predictions, the models have classification tasks, for example, monitoring the severity of Asian rust or target spot in soybeans is a task that demands intensive time and work from qualified professionals to identify the disease and its severity so that there is adequate phytosanitary management. Therefore, the application of spectral data to the SVM model yielded values close to 90% for distinguishing the severities of Asian rust and target spot in soybeans, making it a promising model for identifying the severity of the disease in the field [67,68]. This is useful not only for applications in relation to diseases but also for the identification of pest attacks, such as Red gum lerp psyllid (Glycaspis brimblecombei More, 1964) in eucalyptus, a pest that has been a concern in the cultivation of the crop. It is possible to identify the severity of the attack using spectral information applied to the SVM model, assisting with control management [69]. For example, [70] used eight algorithms to detect attacks by S. frugiperda and D. melacanthus, with Random Forest (RF). The classification of superior soybean genetic materials for industrial grain quality was also possible thanks to the use of ML data with the application of hyperspectral data [71].
Another example of the application of these models is in the prediction of soil nutrients using soil reflectance. In particular, remarkably high levels of accuracy were achieved by the M5P and RF algorithms in predicting soil nutrients, particularly the S content. The authors state that this technology contributes to a reduction in the work involved in the analysis of samples and reduces the use of reagents in laboratory analyses, representing a promising approach for the prediction of soil macronutrients [72]. The use of machine learning has also proven efficient in classifying different soybean genetic materials in the improvement process, specifically in the classification of their primary macronutrient [73,74], secondary macronutrient [75] and micronutrient [76] contents using multispectral reflectance obtained by UAV. This enables the selection of materials with better performance in nutrient absorption.
The application of machine learning algorithms using data from multispectral and hyperspectral sensors represents a promising strategy for monitoring and managing agricultural systems, especially thanks to their ability to detect subtle variations in the physiological state of plants, nutritional and water conditions, soil composition, and other relevant agronomic variables. However, to enable the effective integration of these technologies into real agricultural scenarios, some considerations should be taken into account. First, standardizing data collection and processing protocols is essential, since the heterogeneity between sensors and methodologies limits the comparability and reproducibility of studies. In addition, the construction of robust databases that contemplate different crops, soil and climate conditions, and temporal scales is essential for training models with greater generalization capacity. Additionally, it is essential that the developed models are validated in real field conditions, considering aspects such as spatial variability, spectral resolution, and scalability. Reducing the costs associated with sensors and computational processing, as well as developing more affordable solutions, especially for small and medium-sized producers, should also be prioritized.
5. Challenges of Digital Agriculture
Digital agriculture transforms data collected by sensors into actionable information tailored to farmers’ specific needs. The integration of data from multispectral or hyperspectral sensors, satellite and drone imagery, combined with artificial intelligence, enables the development of systems that mimic human intelligence, training machines to understand and execute specific tasks. Precision agriculture, in turn, utilizes autonomous vehicles to apply fertilizers in precise amounts, reducing waste and optimizing time.
However, digital agriculture still faces several challenges, particularly regarding the accuracy of crop nutritional assessments. One key issue is data variability, which includes the organization of structured and unstructured information. Structured data [77] refers to information organized in specific formats, such as sensor readings, while unstructured data include satellite and drone images that lack a fixed format. It is important to note that although large volumes of data can be difficult to interpret, they provide valuable and robust insights [78].
One way to address this challenge is by generating more comprehensive, detailed, and precise insights. How can this be achieved? The combination of multiple data sources is essential. For instance, in a crop field, agricultural sensor data indicating nutrient deficiencies, such as nitrogen, can be analyzed alongside soil moisture and temperature information. This integrated approach provides a more accurate assessment of field conditions.
This strategy can optimize crop management and nutrient application by adjusting doses and locations based on satellite imagery. Such integration allows for canopy biomass visualization, the identification of plant water stress, and the analysis of crop coverage variability. Plants with lower pigment levels may indicate lower nitrogen content, whereas those with higher pigment levels exhibit increased chlorophyll concentration, resulting in a more intense green coloration and, consequently, higher nitrogen levels. However, it is important to highlight that other nutrient (Mg, Mn, Fe, etc.) are also involved in chlorophyll synthesis, emphasizing the need for accuracy studies to ensure reliable diagnostics.
Based on these data, a predictive mathematical model can be developed to generate fertilizer application maps, optimizing distribution according to the specific needs of different field areas. This approach not only enhances productivity but also reduces fertilizer waste and minimizes groundwater contamination.
6. From Leaf to Harvest: How Interactions Between Spectral Readings, Leaf Nutrient Content, and Yield Affect the Robustness of Predictive Models
Validating plant nutritional diagnostics generated by digital tools, such as spectral analysis, based on actual plant responses in the field is essential to ensuring the advancement of these new techniques on solid foundations. Measuring the accuracy of nutritional diagnostics produced by these technologies through simple field tests is crucial. For instance, if a new method indicates a nutrient deficiency in a specific field, then, theoretically, applying that nutrient should result in a yield increase—provided the diagnosis is correct. To validate this, the nutrient is applied to one part of the field while another portion remains untreated, and the yield of both areas is compared. If the treated area exhibits a yield increase compared to the untreated area, it confirms that the deficiency diagnosis was accurate. Conversely, if no yield improvement is observed, it indicates that the field did not actually suffer from a nutrient deficiency. Field tests to assess the accuracy of new crop nutritional evaluation methods, which rely on leaf tissue analysis to determine nutrient concentrations, have been conducted to verify precision and compare different interpretation methods, such as sufficiency range (conventional method), DRIS/CND, and others. The accuracy of nutritional diagnostics using the conventional method based on leaf tissue analysis has been evaluated in crops such as soybean [79,80,81], banana [82], eucalyptus seedlings [83], and sugarcane [84,85].
However, using crop yield as an indicator of response to validate the accuracy of nutritional status assessments remains limited, even for the conventional methods based on leaf tissue analysis. This is a significant concern. With advancements in digital agriculture, new methods for diagnosing plant nutritional status using spectral techniques have been proposed. Unfortunately, these approaches often fail to incorporate yield data to validate accuracy, which needs to be more widely recognized by researchers in this field.
Accuracy testing through yield analysis is crucial, as it confirms whether an identified deficiency is real by demonstrating that nutrient supplementation results in yield gains. In the study by [44], for example, while the researchers successfully discriminated adequate boron levels in eucalyptus using specific near-infrared bands, the digital tool lacked validation to verify the accuracy of diagnoses in commercial fields using yield data, as previously discussed. Similarly, the authors of [12] estimated the chlorophyll content in various species without characterizing deficient areas or assessing the accuracy of these diagnoses in their study.
Some studies emphasize the correlation between spectral readings and foliar nutrient concentrations as an initial step; however, this does not indicate the accuracy of the generated diagnostics. While nutrient concentration data serve as important reference standards, without accuracy validation tests, they cannot be assumed to be highly reliable. If the conventional method using nutrient concentrations exhibits low diagnostic accuracy, and a new spectral reading method has a high correlation with this conventional approach, it does not necessarily mean the spectral method is highly accurate. This only suggests that the technique can detect the nutrient in plant tissues, which is valuable, but it does not confirm its predictive ability for generating reliable diagnoses.
The authors [47,49,50] studied nutrient application and reported spectral correlations with nutrient concentrations. However, their analysis did not establish the correlation between spectral readings and foliar nutrient content. Even in the study by [46], which presented yield data, there was no conclusive evidence indicating nitrogen deficiency in the studied areas. This highlights the need for assessing “accurate” foliar nutrient levels to confirm such deficiencies.
A critical methodology to address this gap would be the targeted application of nutrients in deficient areas, followed by monitoring the yield response. This approach would allow digital tools to detect nutrient deficiencies that directly impact crop productivity, as well as potential nutrient excesses. Incorporating this methodology would enhance the reliability of spectral analysis-based nutritional diagnostics, making them more applicable to precision agriculture.
A literature search in the Web of Science database using the keywords “deficiency, sensor, fertilization, chlorophyll” retrieved ten scientific articles [84,86,87,88,89,90,91,92]. A similar search in the Scopus database yielded a larger number of relevant studies [93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126]. These findings reinforce that research in this area is advancing, yet the number of high-quality studies remains limited.
In general, studies using spectral reflectance in digital agriculture to generate nutritional diagnostics have demonstrated potential for estimating variations in foliar nutrient concentrations, particularly in controlled-condition experiments. However, the applicability of these findings to field conditions remains debatable. Many researchers in this field mistakenly assume that a strong spectral discrimination associated with increased nutrient levels in the soil or plant is sufficient to ensure that the new tool can generate accurate nutritional diagnostics at the field level. This assumption is problematic because field-grown crops are influenced by numerous environmental and physiological factors that affect foliar nutrient concentrations and, consequently, the accuracy of nutritional diagnostics.
Another critical issue is that most studies focus primarily on nitrogen (N), highlighting the need for research on other essential nutrients. Nitrogen deficiency, for example, may be linked to the insufficient availability of sulfur (S), iron (Fe), or molybdenum (Mo), which play crucial roles in nitrogen metabolism [127].
Additionally, a major limitation of these studies is the lack of accuracy validation using plant response data, specifically yield, to assess the precision of the generated diagnostics. This issue is observed even in conventional foliar analysis methods. The absence of such validation raises concerns about the practical applicability of digital tools in real-world agricultural settings. Without proper field-level validation, these tools may not provide reliable nutritional status assessments for guiding precise fertilization decisions and promoting sustainable agriculture.
To ensure the real-world applicability of spectral reflectance-based diagnostics, complementary studies must be conducted in commercial fields across different regions. This is essential to verify the performance of indirect plant nutritional assessment methods and confirm their reliability in providing accurate diagnostics, ultimately strengthening digital and precision agriculture. Further studies in varied field conditions, using yield data for validation, are considered crucial to confirm the accuracy and reliability of these methods.
7. Final Considerations
Digital agriculture is emerging as a transformative force in crop management, enabling more efficient and sustainable practices. By integrating data from sensors, satellite imagery, and drones with artificial intelligence, precision agriculture becomes more accessible and effective, allowing informed decision-making based on detailed analyses of field conditions. Despite challenges related to data variability and organization, combining multiple information sources provides valuable insights that can optimize crop management and input application.
The recent digital approach, integrating remote sensing technologies, such as spectral sensors on UAVs and satellites, and artificial intelligence, specifically machine learning algorithms like Random Forest, shows potential for improving the precision of predicting leaf nutrient concentrations.
This approach not only enhances productivity but also promotes sustainability by reducing waste and mitigating environmental impacts, such as groundwater contamination. Thus, digital agriculture not only modernizes farming practices but also paves the way for a more responsible and conscious future in food production, emphasizing the importance of a data-driven and integrated approach tailored to the specific needs of each crop.
It is crucial not only to correlate nutrient concentrations with spectral variables but also to investigate how these diagnostics perform in crops under normal, deficient, and excessive nutrient conditions. By gaining a deeper understanding of how these variables impact crop productivity, it will be possible to develop more effective tools with high practical utility in the field—supporting farmers in optimizing plant nutrition and improving yields sustainably.
B.C.d.S. and R.d.M.P.: Conceptualization, Data Curation, Investigation, Project Administration, Resources, Writing—Original Draft, Visualization, Supervision, and Funding Acquisition, Writing—Review and Editing. F.H.R.B.: Visualization, Supervision, Formal Analysis, Conceptualization, Writing—Original Draft, Writing—Review and Editing. C.N.S.C., L.P.R.T., P.E.T. and D.C.S.: Writing—Review and Editing, Writing—Original Draft Preparation, Visualization, Conceptualization. All authors have read and agreed to the published version of the manuscript.
The data used in this study were obtained from scientific articles and other published sources, which are properly cited throughout the manuscript. No new dataset was generated or analyzed for this study.
The authors acknowledge the support from UNESP (Universidade Estadual Paulista) on behalf of the Postgraduate Agronomy Program (Crop Production), CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), UNESP’s Plant Nutrition Study Group (GENPLANT), UFMS (Universidade Federal de Mato Grosso do Sul)—UFMS, and the Cerrado Plant Nutrition Study Group (GECENP).
The authors declare no competing financial interests or known personal relationships that could have influenced the work reported in this article.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Applications of Remote Sensing and Machine Learning for Plant Nutrient and Stress Analysis.
| Nutrient Status/Plant Stress | Sensor and Spectral Bands | Data Type/Visualization | Machine Learning Algorithm/Model | Plant Species | Reference |
|---|---|---|---|---|---|
| Prediction of N, chlorophyll a + b, and carotenoid content | Hyperspectral sensor STS-VIS (Ocean Insight, USA). Spectral bands from 450 to 850 nm | Graphs generated from electromagnetic spectrum readings | Artificial Neural Networks (ANN); M5P decision tree; REPTree (REPT); Random Forest (RF); Polynomial Support Vector Machine (SVMP); ZeroR (ZR) | Maize (Zea mays L.) | [ |
| N, P, and K contents | Reflectance values obtained in the following spectral bands (SBs): red (660 nm), green (550 nm), near-infrared (NIR, 735 nm), and red edge (790 nm) | Spectral images, spectral bands | J48 decision tree and REPTree; Random Forest (RF); Artificial Neural Network (ANN); Support Vector Machine (SVM); Logistic Regression (LR, used as control) | Soybean genotypes (Glycine max L.) | [ |
| Boron deficiency or toxicity | MSI sensor (Sentinel-2) | Orbital images, Vegetation Indices (NDVI, NDRE, EVI, CI, PRSI, CCCI, MTCI, HMSSI) | Spectral Angle Mapping (SAM) algorithm | Ten-year-old clones of Eucalyptus MA-2000 | [ |
| P, K, Mg, Ca, S, B, Zn, Mn, Fe, Cu | 5-band multispectral camera (Micasense Altum, Micasense, USA): blue (465–485 nm), green (550–570 nm), red (663–673 nm), red edge (712–722 nm), and near-infrared (NIR, 820–1000 nm) | Multispectral images from Unmanned Aerial Vehicles (UAVs) | Multiple regression algorithms, such as ElasticNet, Lasso regression, Linear SVM, PLSR, Random Forest (RF), and Ridge regression | Sweet orange seedlings ‘Hamlin’ or ‘Valencia’ grafted onto more than 30 different rootstocks | [ |
| Elevated Nitrogen levels | The Sensefly Sequoia multispectral sensor evaluates vegetation indices in the following spectral bands: green (550 nm), red (660 nm), red edge (735 nm), and near-infrared (NIR, 790 nm) | Image (based on reflectance data), followed by processing of vegetation index models (NDVI, NDRE) | Statistical Model: Yijkl = μ + Bk + Cl + Vi + Nj + CVil + CNjl + VNij + CVNijl + ϵijkl | Maize genotypes (Zea mays L.) | [ |
| N, P, K | Vis-NIR HSI sensor (Visible-Near-Infrared Hyperspectral Imaging) (400–1000 nm) | Visible-Near-Infrared Hyperspectral Image (Vis-NIR HSI) | Partial Least Squares Regression (PLSR) models | Cocoa (Theobroma cacao L.) | [ |
| Predicting leaf nitrogen concentration | Multispectral sensor MicaSense Red-Edge (2017/2018 season) and multispectral sensor Sensefly Parrot Sequoia (2018/2019 season). Bands (G, R, RE, NIR) used for vegetation index calculations, with slight variations between sensors/years (e.g., G: 560/550 nm, R: 668/660 nm, RE: 717/735 nm, NIR: 842/790 nm) | Images and spectral bands | REPTree (REPT); RF; K-Nearest Neighbor (K = 1, K = 5, K = 10) (1NN, 5NN, 10NN); SVM-RBF (SVMR); Polynomial Support Vector Machine (SVMP); Linear Regression (LR); Radial Basis Function Regression (RBF) | 11 Maize cultivars | [ |
| Predicting nitrogen content | Multispectral sensor Parrot Sequoia. Spectral bands captured: green (530–570 nm), red (640–680 nm), red edge (730–740 nm), and near-infrared (NIR, 770–810 nm) | Multispectral images | RF; ANN; SVM; Decision Trees (DT); RNA (Artificial Neural Network—same as ANN) | 33,600 Valencia orange trees (Citrus sinensis ‘Valencia’) planted on Citrumelo oscillating rootstock | [ |
| Predicting leaf nitrogen concentration: low (≤27 g·kg−1), medium (>27 & ≤29 g·kg−1), and high (>29 g·kg−1) | Parrot Sequoia sensor. Green (510–590 nm), red (620–700 nm), red edge (725–745 nm), and near-infrared (NIR, 750–830 nm) bands | Multispectral images, spectral bands | Constrained Energy Minimization; Linear Spectral Unmixing; Mixture-Tuned Matched Filtering; Minimum Distance; Orthogonal Subspace Projection; Spectral Angle Mapper; 1 Spectral Information Divergence | Valencia sweet orange (Citrus sinensis ‘Valencia’), grafted onto Citrumelo Swingle rootstock | [ |
Abbreviations used in the table: ANN: Artificial Neural Networks; RF: Random Forest; SVM: Support Vector Machine; PLSR: Partial Least Squares Regression; REPT/REPTree: Reduced Error Pruning Tree; LR: Logistic Regression; MSI: Multispectral Imager; HSI: Hyperspectral Imaging; NIR: Near-Infrared; NDVI: Normalized Difference Vegetation Index; NDRE: Normalized Difference Red Edge Index; UAV: Unmanned Aerial Vehicle.
1. Sharma, A.; Prakash, A.; Bhambota, S.; Kumar, S. Investigations of Precision Agriculture Technologies with Application to Developing Countries. Environ. Dev. Sustain.; 2024; 26, 131. [DOI: https://dx.doi.org/10.1007/s10668-024-04572-y]
2. Gebbers, R.; Adamchuk, V.I. Precision Agriculture and Food Security. Science; 2010; 327, pp. 828-831. [DOI: https://dx.doi.org/10.1126/science.1183899] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20150492]
3. Himesh, S.; Rao, E.V.S.P.; Gouda, K.C.; Ramesh, K.V.; Rakesh, V.; Mohapatra, G.N.; Rao, B.K.; Sahoo, S.K.; Ajilesh, P. Digital Revolution and Big Data: A New Revolution in Agriculture. CABI Rev.; 2018; pp. 1-7. [DOI: https://dx.doi.org/10.1079/PAVSNNR201813021]
4. Zhang, N.; Wang, M.; Wang, N. Precision Agriculture—A Worldwide Overview. Comput. Electron. Agric.; 2002; 36, pp. 113-132. [DOI: https://dx.doi.org/10.1016/S0168-1699(02)00096-0]
5. Teodoro, L.P.R.; Estevão, R.; Santana, D.C.; de Oliveira, I.C.; Lopes, M.T.G.; de Azevedo, G.B.; Rojo Baio, F.H.; da Silva Junior, C.A.; Teodoro, P.E. Eucalyptus Species Discrimination Using Hyperspectral Sensor Data and Machine Learning. Forests; 2023; 15, 39. [DOI: https://dx.doi.org/10.3390/f15010039]
6. Bhakthan, S.M.; Loganathan, A. A Hyperspectral Unmixing Model Using Convolutional Vision Transformer. Earth Sci. Inform.; 2024; 17, pp. 2255-2273. [DOI: https://dx.doi.org/10.1007/s12145-024-01273-y]
7. Comba, L.; Biglia, A.; Ricauda Aimonino, D.; Tortia, C.; Mania, E.; Guidoni, S.; Gay, P. Leaf Area Index Evaluation in Vineyards Using 3D Point Clouds from UAV Imagery. Precis. Agric.; 2020; 21, pp. 881-896. [DOI: https://dx.doi.org/10.1007/s11119-019-09699-x]
8. Abiri, R.; Rizan, N.; Balasundram, S.K.; Shahbazi, A.B.; Abdul-Hamid, H. Application of Digital Technologies for Ensuring Agricultural Productivity. Heliyon; 2023; 9, e22601. [DOI: https://dx.doi.org/10.1016/j.heliyon.2023.e22601] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38125472]
9. Teodoro, P.E.; Rossi, F.S.; Teodoro, L.P.R.; Santana, D.C.; Ratke, R.F.; de Oliveira, I.C.; Della Silva, J.L.; de Oliveira, J.L.G.; da Silva, N.P.; Baio, F.H.R.
10. Xie, K.; Cakmak, I.; Wang, S.; Zhang, F.; Guo, S. Synergistic and Antagonistic Interactions between Potassium and Magnesium in Higher Plants. Crop J.; 2021; 9, pp. 249-256. [DOI: https://dx.doi.org/10.1016/j.cj.2020.10.005]
11. da Silva, B.C.; Prado, R.d.M.; Baio, F.H.R.; Campos, C.N.S.; Teodoro, L.P.R.; Teodoro, P.E.; Santana, D.C.; Fernandes, T.F.S.; da Silva Junior, C.A.; Loureiro, E.d.S. New Approach for Predicting Nitrogen and Pigments in Maize from Hyperspectral Data and Machine Learning Models. Remote Sens. Appl.; 2024; 33, 101110. [DOI: https://dx.doi.org/10.1016/j.rsase.2023.101110]
12. Gitelson, A.A.; Gritz, Y.; Merzlyak, M.N. Relationships between Leaf Chlorophyll Content and Spectral Reflectance and Algorithms for Non-Destructive Chlorophyll Assessment in Higher Plant Leaves. J. Plant Physiol.; 2003; 160, pp. 271-282. [DOI: https://dx.doi.org/10.1078/0176-1617-00887]
13. Peng, Y.; Zhang, M.; Xu, Z.; Yang, T.; Su, Y.; Zhou, T.; Wang, H.; Wang, Y.; Lin, Y. Estimation of Leaf Nutrition Status in Degraded Vegetation Based on Field Survey and Hyperspectral Data. Sci. Rep.; 2020; 10, 4361. [DOI: https://dx.doi.org/10.1038/s41598-020-61294-7]
14. Song, G.; Wang, Q. Developing Hyperspectral Indices for Assessing Seasonal Variations in the Ratio of Chlorophyll to Carotenoid in Deciduous Forests. Remote Sens.; 2022; 14, 1324. [DOI: https://dx.doi.org/10.3390/rs14061324]
15. Anastasiou, E.; Fountas, S.; Voulgaraki, M.; Psiroukis, V.; Koutsiaras, M.; Kriezi, O.; Lazarou, E.; Vatsanidou, A.; Fu, L.; Bartolo, F.D.
16. Wei, H.; Xu, W.; Kang, B.; Eisner, R.; Muleke, A.; Rodriguez, D.; deVoil, P.; Sadras, V.; Monjardino, M.; Harrison, M.T. Irrigation with Artificial Intelligence: Problems, Premises, Promises. Hum.-Centric Intell. Syst.; 2024; 4, pp. 187-205. [DOI: https://dx.doi.org/10.1007/s44230-024-00072-4]
17. Ahmed, Z.; Wan, S.; Zhang, F.; Zhong, W. Artificial Intelligence for Omics Data Analysis. BMC Methods; 2024; 1, 4. [DOI: https://dx.doi.org/10.1186/s44330-024-00004-5]
18. Holl, C. The Content Intelligence: An Argument against the Lethality of Artificial Intelligence. Discov. Artif. Intell.; 2024; 4, 13. [DOI: https://dx.doi.org/10.1007/s44163-024-00112-9]
19. Gignac, G.E.; Szodorai, E.T. Defining Intelligence: Bridging the Gap between Human and Artificial Perspectives. Intelligence; 2024; 104, 101832. [DOI: https://dx.doi.org/10.1016/j.intell.2024.101832]
20. Prasad, G.A.; Kumar, A.V.S.; Sharma, P.; Irawati, I.D.; Chandrashekar, D.V.; Musirin, I.B.; Abdullah, H.M.A.; Rao, L.M. Artificial Intelligence in Computer Science; IGI Global: Hershey, PA, USA, 2023; pp. 43-60.
21. Gottfredson, L.S. Mainstream Science on Intelligence. Wall Street J.; 1997.
22. Fabbri, A.G.; Fung, K.B.; Yatabe, S.M. The Usage of Artificial Intelligence in Remote Sensing: A Review of Applications and Current Research. Quantitative Analysis of Mineral and Energy Resources; Springer: Dordrecht, The Netherlands, 1988; pp. 489-512.
23. Akosah, S.; Gratchev, I.; Kim, D.-H.; Ohn, S.-Y. Application of Artificial Intelligence and Remote Sensing for Landslide Detection and Prediction: Systematic Review. Remote Sens.; 2024; 16, 2947. [DOI: https://dx.doi.org/10.3390/rs16162947]
24. Zhang, Q. Opinion Paper: Precision Agriculture, Smart Agriculture, or Digital Agriculture. Comput. Electron. Agric.; 2023; 211, 107982. [DOI: https://dx.doi.org/10.1016/j.compag.2023.107982]
25. Marconi, S.; Weinstein, B.G.; Zou, S.; Bohlman, S.A.; Zare, A.; Singh, A.; Stewart, D.; Harmon, I.; Steinkraus, A.; White, E.P. Continental-Scale Hyperspectral Tree Species Classification in the United States National Ecological Observatory Network. Remote Sens. Environ.; 2022; 282, 113264. [DOI: https://dx.doi.org/10.1016/j.rse.2022.113264]
26. Katharria, A.; Rajwar, K.; Pant, M.; Velásquez, J.D.; Snášel, V.; Deep, K. Information Fusion in Smart Agriculture: Machine Learning Applications and Future Research Directions. SSRN; 2025; [DOI: https://dx.doi.org/10.2139/ssrn.5187083]
27. Padhiary, M.; Saha, D.; Kumar, R.; Sethi, L.N.; Kumar, A. Enhancing precision agriculture: A comprehensive review of machine learning and AI vision applications in all-terrain vehicle for farm automation. Smart Agric. Technol.; 2024; 8, 100483. [DOI: https://dx.doi.org/10.1016/j.atech.2024.100483]
28. Addison, M.; Bonuedi, I.; Arhin, A.A.; Wadei, B.; Owusu-Addo, E.; Fredua Antoh, E.; Mensah-Odum, N. Exploring the Impact of Agricultural Digitalization on Smallholder Farmers’ Livelihoods in Ghana. Heliyon; 2024; 10, e27541. [DOI: https://dx.doi.org/10.1016/j.heliyon.2024.e27541]
29. Fossen, F.; Sorgner, A. Mapping the Future of Occupations: Transformative and Destructive Effects of New Digital Technologies on Jobs. Foresight STI Gov.; 2019; 13, pp. 10-18. [DOI: https://dx.doi.org/10.17323/2500-2597.2019.2.10.18]
30. Klerkx, L.; Jakku, E.; Labarthe, P. A Review of Social Science on Digital Agriculture, Smart Farming and Agriculture 4.0: New Contributions and a Future Research Agenda. NJAS Wagening J. Life Sci.; 2019; 90, pp. 1-16. [DOI: https://dx.doi.org/10.1016/j.njas.2019.100315]
31. ONU. Objetivos de Desenvolvimento Sustentável: As Nações Unidas No Brasil; Organização das Nações Unidas: San Francisco, CA, USA, 2021.
32. MacPherson, J.; Voglhuber-Slavinsky, A.; Olbrisch, M.; Schöbel, P.; Dönitz, E.; Mouratiadou, I.; Helming, K. Future Agricultural Systems and the Role of Digitalization for Achieving Sustainability Goals. A Review. Agron. Sustain. Dev.; 2022; 42, 70. [DOI: https://dx.doi.org/10.1007/s13593-022-00792-6]
33. Balkrishna, A.; Pathak, R.; Kumar, S.; Arya, V.; Singh, S.K. A Comprehensive Analysis of the Advances in Indian Digital Agricultural Architecture. Smart Agric. Technol.; 2023; 5, 100318. [DOI: https://dx.doi.org/10.1016/j.atech.2023.100318]
34. Chandra, R.; Collis, S. Digital Agriculture for Small-Scale Producers. Commun. ACM; 2021; 64, pp. 75-84. [DOI: https://dx.doi.org/10.1145/3454008]
35. Abioye, E.A.; Hensel, O.; Esau, T.J.; Elijah, O.; Abidin, M.S.Z.; Ayobami, A.S.; Yerima, O.; Nasirahmadi, A. Precision Irrigation Management Using Machine Learning and Digital Farming Solutions. AgriEngineering; 2022; 4, pp. 70-103. [DOI: https://dx.doi.org/10.3390/agriengineering4010006]
36. Ogubuike, R.; Adib, A.; Orji, R. Masa: AI-Adaptive Mobile App for Sustainable Agriculture. Proceedings of the 2021 IEEE 12th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON); Vancouver, BC, Canada, 27–30 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 1064-1069.
37. Basso, B.; Antle, J. Digital Agriculture to Design Sustainable Agricultural Systems. Nat. Sustain.; 2020; 3, pp. 254-256. [DOI: https://dx.doi.org/10.1038/s41893-020-0510-0]
38. Abbasi, R.; Martinez, P.; Ahmad, R. The Digitization of Agricultural Industry—A Systematic Literature Review on Agriculture 4.0. Smart Agric. Technol.; 2022; 2, 100042. [DOI: https://dx.doi.org/10.1016/j.atech.2022.100042]
39. O’Donoghue, T.; Minasny, B.; McBratney, A. Digital Regenerative Agriculture. npj Sustain. Agric.; 2024; 2, 5. [DOI: https://dx.doi.org/10.1038/s44264-024-00012-6]
40. McBratney, A.; Whelan, B.; Ancev, T.; Bouma, J. Future Directions of Precision Agriculture. Precis. Agric.; 2005; 6, pp. 7-23. [DOI: https://dx.doi.org/10.1007/s11119-005-0681-8]
41. Porciello, J.; Coggins, S.; Mabaya, E.; Otunba-Payne, G. Digital Agriculture Services in Low- and Middle-Income Countries: A Systematic Scoping Review. Glob. Food Secur.; 2022; 34, 100640. [DOI: https://dx.doi.org/10.1016/j.gfs.2022.100640]
42. Mentsiev, A.U.; Amirova, E.F.; Afanasev, N. V Digitalization and Mechanization in Agriculture Industry. IOP Conf. Ser. Earth Environ. Sci.; 2020; 548, 032031. [DOI: https://dx.doi.org/10.1088/1755-1315/548/3/032031]
43. Santana, D.C.; Teixeira Filho, M.C.M.; da Silva, M.R.; das Chagas, P.H.M.; de Oliveira, J.L.G.; Baio, F.H.R.; Campos, C.N.S.; Teodoro, L.P.R.; da Silva Junior, C.A.; Teodoro, P.E.
44. Damasceno, A.S.d.S.; Boechat, C.L.; de Souza, H.A.; Capristo-Silva, G.F.; Mendes, W.D.S.; Teodoro, P.E.; Morais, P.G.C.; de Oliveira, R.I.; Della-Silva, J.L.; de Souza, I.A.M.
45. Costa, L.; Kunwar, S.; Ampatzidis, Y.; Albrecht, U. Determining Leaf Nutrient Concentrations in Citrus Trees Using UAV Imagery and Machine Learning. Precis. Agric.; 2022; 23, pp. 854-875. [DOI: https://dx.doi.org/10.1007/s11119-021-09864-1]
46. Santana, D.C.; Cotrim, M.F.; Flores, M.S.; Rojo Baio, F.H.; Shiratsuchi, L.S.; Antonio da Silva Junior, C.; Ribeiro Teodoro, L.P.; Teodoro, P.E. UAV-BASED Multispectral Sensor to Measure Variations in Corn as a Function of Nitrogen Topdressing. Remote Sens. Appl.; 2021; 23, 100534. [DOI: https://dx.doi.org/10.1016/j.rsase.2021.100534]
47. Malmir, M.; Tahmasbian, I.; Xu, Z.; Farrar, M.B.; Bai, S.H. Prediction of Macronutrients in Plant Leaves Using Chemometric Analysis and Wavelength Selection. J. Soils Sediments; 2020; 20, pp. 249-259. [DOI: https://dx.doi.org/10.1007/s11368-019-02418-z]
48. Osco, L.P.; Marques Ramos, A.P.; Saito Moriya, É.A.; de Souza, M.; Marcato Junior, J.; Matsubara, E.T.; Imai, N.N.; Creste, J.E. Improvement of Leaf Nitrogen Content Inference in Valencia-Orange Trees Applying Spectral Analysis Algorithms in UAV Mounted-Sensor Images. Int. J. Appl. Earth Obs. Geoinf.; 2019; 83, 101907. [DOI: https://dx.doi.org/10.1016/j.jag.2019.101907]
49. Osco, L.P.; Marques Ramos, A.P.; Roberto Pereira, D.; Akemi Saito Moriya, É.; Nobuhiro Imai, N.; Takashi Matsubara, E.; Estrabis, N.; de Souza, M.; Marcato Junior, J.; Gonçalves, W.N.
50. Osco, L.P.; Arruda, M.d.S.; Junior, J.M.; Silva, N.B.d.; Ramos, A.P.M.; Moryia, É.A.S.; Imai, N.N.; Pereira, D.R.; Creste, J.E.; Matsubara, E.T.
51. Quaggio, J.A.; Cantarella, H.; van Raij, B. Phosphorus and Potassium Soil Test and Nitrogen Leaf Analysis as a Base for Citrus Fertilization. Nutr. Cycl. Agroecosyst.; 1998; 52, pp. 67-74. [DOI: https://dx.doi.org/10.1023/A:1009763027607]
52. Goodfellow, I.; Bengio, Y.; Courville, A. Deep Learning; MIT Press: Cambridge, MA, USA, 2016.
53. Haykin, S.S. Neural Networks and Learning Machines; Prentice Hall/Pearson: London, UK, 2009; ISBN 9780131471399
54. Quinlan, J.R. Learning with Continuous Classes. Proceedings of the 5th Australian Joint Conference on Artificial Intelligence; Hobart, Australia, 16–18 November 1992; Volume 92, pp. 343-348. [DOI: https://dx.doi.org/10.1016/S0020-7373(87)80053-6]
55. Pande, C.B.; Al-Ansari, N.; Kushwaha, N.L.; Srivastava, A.; Noor, R.; Kumar, M.; Moharir, K.N.; Elbeltagi, A. Forecasting of SPI and Meteorological Drought Based on the Artificial Neural Network and M5P Model Tree. Land; 2022; 11, 2040. [DOI: https://dx.doi.org/10.3390/land11112040]
56. Sihag, P.; Mehta, T.; Sammen, S.S.; Pande, C.B.; Puri, D.; Radwan, N. Predictive Modelling of Nitrogen Dioxide Using Soft Computing Techniques in the Agra, Uttar Pradesh, India. Phys. Chem. Earth Parts A/B/C; 2024; 134, 103589. [DOI: https://dx.doi.org/10.1016/j.pce.2024.103589]
57. Abellán-García, J.; Guzmán-Guzmán, J.S. Random Forest-Based Optimization of UHPFRC under Ductility Requirements for Seismic Retrofitting Applications. Constr. Build. Mater.; 2021; 285, 122869. [DOI: https://dx.doi.org/10.1016/j.conbuildmat.2021.122869]
58. Breiman, L. Random Forests. Mach. Learn.; 2001; 45, pp. 5-32. [DOI: https://dx.doi.org/10.1023/A:1010933404324]
59. Auret, L.; Aldrich, C. Interpretation of Nonlinear Relationships between Process Variables by Use of Random Forests. Miner. Eng.; 2012; 35, pp. 27-42. [DOI: https://dx.doi.org/10.1016/j.mineng.2012.05.008]
60. Guido, R.; Ferrisi, S.; Lofaro, D.; Conforti, D. An Overview on the Advancements of Support Vector Machine Models in Healthcare Applications: A Review. Information; 2024; 15, 235. [DOI: https://dx.doi.org/10.3390/info15040235]
61. Pathak, D.K.; Kalita, S.K.; Bhattacharya, D.K. Hyperspectral Image Classification Using Support Vector Machine: A Spectral Spatial Feature Based Approach. Evol. Intell.; 2022; 15, pp. 1809-1823. [DOI: https://dx.doi.org/10.1007/s12065-021-00591-0]
62. Witten, I.H.; Frank, E. Data Mining: Practical Machine Learning Tools and Techniques; 2nd ed. Elseiver: Amsterdam, The Netherlands, 2005.
63. Santana, D.C.; Ratke, R.F.; Zanatta, F.L.; Campos, C.N.S.; Seron, A.C.d.S.C.; Teodoro, L.P.R.; Silva, N.P.d.; Oliveira, G.S.; dos Santos, R.G.; Alvarez, R.d.C.F. Caffeine Content Prediction in Coffee Beans Using Hyperspectral Reflectance and Machine Learning. AgriEngineering; 2024; 6, pp. 4480-4492. [DOI: https://dx.doi.org/10.3390/agriengineering6040255]
64. Osco, L.P.; Junior, J.M.; Ramos, A.P.; Furuya, D.E.; Santana, D.C.; Teodoro, L.P.; Gonçalves, W.N.; Baio, F.H.; Pistori, H.; Junior, C.A.
65. Kumar, C.; Dhillon, J.; Huang, Y.; Reddy, K. Explainable machine learning models for corn yield prediction using UAV multispectral data. Comput. Electron. Agric.; 2025; 231, 109990. [DOI: https://dx.doi.org/10.1016/j.compag.2025.109990]
66. Manfroi Filho, E.A.; Teodoro, P.E.; Teodoro, L.P.R.; Santana, D.C.; Azevedo, G.T.d.O.S.; da Silva Junior, C.A.; Baio, F.H.R.; de Azevedo, G.B. VNIR-SWIR Spectroscopy and Machine Learning for Measuring Dendrometric Variables in Native Species. Remote Sens. Appl.; 2025; 37, 101522. [DOI: https://dx.doi.org/10.1016/j.rsase.2025.101522]
67. Santana, D.C.; de Queiroz Otone, J.D.; Baio, F.H.R.; Teodoro, L.P.R.; Alves, M.E.M.; da Silva Junior, C.A.; Teodoro, P.E. Machine Learning in the Classification of Asian Rust Severity in Soybean Using Hyperspectral Sensor. Spectrochim. Acta A Mol. Biomol. Spectrosc.; 2024; 313, 124113. [DOI: https://dx.doi.org/10.1016/j.saa.2024.124113]
68. de Queiroz Otone, J.D.; Theodoro, G.d.F.; Santana, D.C.; Teodoro, L.P.R.; de Oliveira, J.T.; de Oliveira, I.C.; da Silva Junior, C.A.; Teodoro, P.E.; Baio, F.H.R. Hyperspectral Response of the Soybean Crop as a Function of Target Spot (Corynespora cassiicola) Using Machine Learning to Classify Severity Levels. AgriEngineering; 2024; 6, pp. 330-343. [DOI: https://dx.doi.org/10.3390/agriengineering6010020]
69. de Gregori, G.S.; de Souza Loureiro, E.; Amorim Pessoa, L.G.; Azevedo, G.B.d.; Azevedo, G.T.d.O.S.; Santana, D.C.; de Oliveira, I.C.; de Oliveira, J.L.G.; Teodoro, L.P.R.; Baio, F.H.R. Machine Learning in the Hyperspectral Classification of Glycaspis Brimblecombei (Hemiptera psyllidae) Attack Severity in Eucalyptus. Remote Sens.; 2023; 15, 5657. [DOI: https://dx.doi.org/10.3390/rs15245657]
70. Furuya, D.E.G.; Ma, L.; Pinheiro, M.M.F.; Gomes, F.D.G.; Gonçalvez, W.N.; Junior, J.M.; de Castro Rodrigues, D.; Blassioli-Moraes, M.C.; Michereff, M.F.F.; Borges, M. Prediction of Insect-Herbivory-Damage and Insect-Type Attack in Maize Plants Using Hyperspectral Data. Int. J. Appl. Earth Obs. Geoinf.; 2021; 105, 102608.
71. Bernardy, R.; Cañizares, L.d.C.C.; Meza, S.L.R.; Rodrigues, L.A.; Jappe, S.N.; Oliveira, M.d. Identifying soybean genotypes with artificial intelligence and near infrared reflectance. Cienc. e Agrotecnologia; 2024; 48, e005224. [DOI: https://dx.doi.org/10.1590/1413-7054202448005224]
72. Nozari, S.; Pahlavan-Rad, M.R.; Brungard, C.; Heung, B.; Borůvka, L. Digital soil mapping using machine learning-based methods to predict soil organic carbon in two different districts in the Czech Republic. Soil Water Res.; 2024; 19, pp. 32-49. [DOI: https://dx.doi.org/10.17221/119/2023-SWR]
73. Mukherjee, S.; Laskar, S. Vis–NIR-Based Optical Sensor System for Estimation of Primary Nutrients in Soil. J. Opt.; 2019; 48, pp. 87-103. [DOI: https://dx.doi.org/10.1007/s12596-019-00517-1]
74. Zhou, J.; Beche, E.; Vieira, C.C.; Yungbluth, D.; Zhou, J.; Scaboo, A.; Chen, P. Improve Soybean Variety Selection Accuracy Using UAV-Based High-Throughput Phenotyping Technology. Front. Plant Sci.; 2022; 12, 768742. [DOI: https://dx.doi.org/10.3389/fpls.2021.768742]
75. Santana, D.C.; de Oliveira, I.C.; Cavalheiro, S.B.; das Chagas, P.H.M.; Teixeira Filho, M.C.M.; Della-Silva, J.L.; Teodoro, L.P.R.; Campos, C.N.S.; Baio, F.H.R.; da Silva Junior, C.A.
76. Keshavarzi, A.; Kaya, F.; Başayiğit, L.; Gyasi-Agyei, Y.; Rodrigo-Comino, J.; Caballero-Calvo, A. Spatial prediction of soil micronutrients using machine learning algorithms integrated with multiple digital covariates. Nutr. Cycl. Agroecosyst.; 2023; 127, pp. 137-153. [DOI: https://dx.doi.org/10.1007/s10705-023-10303-y]
77. Lowenberg-DeBoer, J.; Huang, I.Y.; Grigoriadis, V.; Blackmore, S. Economics of Robots and Automation in Field Crop Production. Precis. Agric.; 2020; 21, pp. 278-299. [DOI: https://dx.doi.org/10.1007/s11119-019-09667-5]
78. Zhang, Q.; Yang, L.T.; Chen, Z.; Li, P. A Survey on Deep Learning for Big Data. Inf. Fusion; 2018; 42, pp. 146-157. [DOI: https://dx.doi.org/10.1016/j.inffus.2017.10.006]
79. Beverly, R.B. DRIS Diagnoses of Soybean Nitrogen, Phosphorus, and Potassium Status Are Unsatisfactory. J. Plant Nutr.; 1993; 16, pp. 1431-1447. [DOI: https://dx.doi.org/10.1080/01904169309364625]
80. Traspadini, E.I.F.; Wadt, P.G.S.; de Mello-Prado, R.; Garcia-Roque, C.; Roberto-Wassolowski, C.; Vidal-Perez, D. Efficiency of Critical Level and Compositional Nutrient Diagnosis Methods to Evaluate Boron Nutritional Status in Soybean. Chil. J. Agric. Res.; 2022; 82, pp. 309-319. [DOI: https://dx.doi.org/10.4067/S0718-58392022000200309]
81. Traspadini, E.I.F.; Wadt, P.G.S.; de Mello Prado, R.; Roque, C.G.; Wassolowski, C.R. Prescient Diagnostic Analysis for Boron Nutritional Status in Soy Crops. Sci. Rep.; 2023; 13, 2281. [DOI: https://dx.doi.org/10.1038/s41598-022-26263-2] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36759536]
82. Villaseñor-Ortiz, D.; de Mello Prado, R.; Pereira da Silva, G.; Lata-Tenesaca, L.F. Applicability of DRIS in Bananas Based on the Accuracy of Nutritional Diagnoses for Nitrogen and Potassium. Sci. Rep.; 2022; 12, 18125. [DOI: https://dx.doi.org/10.1038/s41598-022-22554-w]
83. de Morais, T.C.B.; Prado, R.d.M.; Traspadini, E.I.F.; Wadt, P.G.S.; de Paula, R.C.; Rocha, A.M.S. Efficiency of the CL, DRIS and CND Methods in Assessing the Nutritional Status of Eucalyptus Spp. Rooted Cuttings. Forests; 2019; 10, 786. [DOI: https://dx.doi.org/10.3390/f10090786]
84. Pereira da Silva, G.; Prado, R.d.M.; Wadt, P.G.S.; da Silva, F.C.; Vaz, G.J.; Moda, L.R.; Caione, G. Modeling Formulas of the Comprehensive Diagnosis and Recommendation System (DRIS) for Phosphorus in Sugarcane. J. Plant Nutr.; 2021; 44, pp. 1316-1329. [DOI: https://dx.doi.org/10.1080/01904167.2020.1862192]
85. da Silva, G.P.; Wadt, P.G.S.; Prado, R.d.M.; Caione, G.; Moda, L.R. Accuracy of Plant Response Potential to Fertilization in Nutritional Diagnoses for Phosphorus in Sugarcane. J. Plant Nutr.; 2022; 45, pp. 1702-1711. [DOI: https://dx.doi.org/10.1080/01904167.2021.2014869]
86. Barker, D.W.; Sawyer, J.E. Using Active Canopy Sensing to Adjust Nitrogen Application Rate in Corn. Agron. J.; 2012; 104, pp. 926-933. [DOI: https://dx.doi.org/10.2134/agronj2012.0030]
87. Corp, L.A.; McMurtrey, J.E.; Middleton, E.M.; Mulchi, C.L.; Chappelle, E.W.; Daughtry, C.S.T. Fluorescence Sensing Systems: In Vivo Detection of Biophysical Variations in Field Corn Due to Nitrogen Supply. Remote Sens. Environ.; 2003; 86, pp. 470-479. [DOI: https://dx.doi.org/10.1016/S0034-4257(03)00125-1]
88. Motomiya, A.V.d.A.; Molin, J.P.; Motomiya, W.R.; Biscaro, G.A. Diagnose Nutricional Com o Uso de Sensor Óptico Ativo Em Algodoeiro. Rev. Bras. de Eng. Agrícola e Ambient.; 2012; 16, pp. 1159-1165. [DOI: https://dx.doi.org/10.1590/S1415-43662012001100003]
89. Padilla, F.M.; Farneselli, M.; Gianquinto, G.; Tei, F.; Thompson, R.B. Monitoring Nitrogen Status of Vegetable Crops and Soils for Optimal Nitrogen Management. Agric. Water Manag.; 2020; 241, 106356. [DOI: https://dx.doi.org/10.1016/j.agwat.2020.106356]
90. Pancorbo, J.L.; Camino, C.; Alonso-Ayuso, M.; Raya-Sereno, M.D.; Gonzalez-Fernandez, I.; Gabriel, J.L.; Zarco-Tejada, P.J.; Quemada, M. Simultaneous Assessment of Nitrogen and Water Status in Winter Wheat Using Hyperspectral and Thermal Sensors. Eur. J. Agron.; 2021; 127, 126287. [DOI: https://dx.doi.org/10.1016/j.eja.2021.126287]
91. Quemada, M.; Gabriel, J.; Zarco-Tejada, P. Airborne Hyperspectral Images and Ground-Level Optical Sensors As Assessment Tools for Maize Nitrogen Fertilization. Remote Sens.; 2014; 6, pp. 2940-2962. [DOI: https://dx.doi.org/10.3390/rs6042940]
92. Wang, Y.; Dunn, B.L.; Arnall, D.B. Assessing Nitrogen Status in Potted Geranium through Discriminant Analysis of Ground-Based Spectral Reflectance Data. HortScience; 2012; 47, pp. 343-348. [DOI: https://dx.doi.org/10.21273/HORTSCI.47.3.343]
93. Adams, M.L.; Norvell, W.A.; Philpot, W.D.; Peverly, J.H. Spectral Detection of Micronutrient Deficiency in ‘Bragg’ Soybean. Agron. J.; 2000; 92, pp. 261-268. [DOI: https://dx.doi.org/10.2134/agronj2000.922261x]
94. Avgoustaki, D.D.; Avgoustakis, I.; Miralles, C.C.; Sohn, J.; Xydis, G. Autonomous Mobile Robot with Attached Multispectral Camera to Monitor the Development of Crops and Detect Nutrient and Water Deficiencies in Vertical Farms. Agronomy; 2022; 12, 2691. [DOI: https://dx.doi.org/10.3390/agronomy12112691]
95. Ayala-Silva, T.; Beyl, C.A. Changes in Spectral Reflectance of Wheat Leaves in Response to Specific Macronutrient Deficiency. Adv. Space Res.; 2005; 35, pp. 305-317. [DOI: https://dx.doi.org/10.1016/j.asr.2004.09.008]
96. Bhadra, S.; Sagan, V.; Maimaitijiang, M.; Maimaitiyiming, M.; Newcomb, M.; Shakoor, N.; Mockler, T.C. Quantifying Leaf Chlorophyll Concentration of Sorghum from Hyperspectral Data Using Derivative Calculus and Machine Learning. Remote Sens.; 2020; 12, 2082. [DOI: https://dx.doi.org/10.3390/rs12132082]
97. Campbell, P.K.E.; Middleton, E.M.; Corp, L.A.; Kim, M.S. Contribution of Chlorophyll Fluorescence to the Apparent Vegetation Reflectance. Sci. Total Environ.; 2008; 404, pp. 433-439. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2007.11.004]
98. Cartelat, A.; Cerovic, Z.G.; Goulas, Y.; Meyer, S.; Lelarge, C.; Prioul, J.-L.; Barbottin, A.; Jeuffroy, M.-H.; Gate, P.; Agati, G.
99. Crema, A.; Boschetti, M.; Nutini, F.; Cillis, D.; Casa, R. Influence of Soil Properties on Maize and Wheat Nitrogen Status Assessment from Sentinel-2 Data. Remote Sens.; 2020; 12, 2175. [DOI: https://dx.doi.org/10.3390/rs12142175]
100. Gomes, R.L.; Sousa, M.C.; Campos, F.G.; Boaro, C.S.F.; de Souza Passos, J.R.; Ferreira, G. Near-Infrared Leaf Reflectance Modeling of Annona Emarginata Seedlings for Early Detection of Variations in Nitrogen Concentration. J. For. Res.; 2023; 34, pp. 269-282. [DOI: https://dx.doi.org/10.1007/s11676-022-01557-3]
101. Hong, S.; Schepers, J.S.; Francis, D.D.; Schlemmer, M.R. Comparison of Ground-Based Remote Sensors for Evaluation of Corn Biomass Affected by Nitrogen Stress. Commun. Soil Sci. Plant Anal.; 2007; 38, pp. 2209-2226. [DOI: https://dx.doi.org/10.1080/00103620701549157]
102. Hunt, E.R.; Horneck, D.A.; Spinelli, C.B.; Turner, R.W.; Bruce, A.E.; Gadler, D.J.; Brungardt, J.J.; Hamm, P.B. Monitoring Nitrogen Status of Potatoes Using Small Unmanned Aerial Vehicles. Precis. Agric.; 2018; 19, pp. 314-333. [DOI: https://dx.doi.org/10.1007/s11119-017-9518-5]
103. Jang, S.; Han, J.; Cho, J.; Jung, J.; Lee, S.; Lee, D.; Kim, J. Estimation of Apple Leaf Nitrogen Concentration Using Hyperspectral Imaging-Based Wavelength Selection and Machine Learning. Horticulturae; 2023; 10, 35. [DOI: https://dx.doi.org/10.3390/horticulturae10010035]
104. Kruse, J.K.; Christians, N.E.; Chaplin, M.H. Remote Sensing of Nitrogen Stress in Creeping Bentgrass. Agron. J.; 2006; 98, pp. 1640-1645. [DOI: https://dx.doi.org/10.2134/agronj2006.0022]
105. Lorenzo, G.A.; Mascarini, L.; Gonzalez, M.N. Dosis de N Sobre Reflectancia Espectral, Contenido de Clorofila y Nutrientes En Plantas de Gerbera. Hortic. Bras.; 2017; 35, pp. 278-285. [DOI: https://dx.doi.org/10.1590/s0102-053620170220]
106. Maresma, Á.; Lloveras, J.; Martínez-Casasnovas, J. Use of Multispectral Airborne Images to Improve In-Season Nitrogen Management, Predict Grain Yield and Estimate Economic Return of Maize in Irrigated High Yielding Environments. Remote Sens.; 2018; 10, 543. [DOI: https://dx.doi.org/10.3390/rs10040543]
107. Mariotti, M.; Ercoli, L.; Masoni, A. Spectral Properties of Iron-Deficient Corn and Sunflower Leaves. Remote Sens. Environ.; 1996; 58, pp. 282-288. [DOI: https://dx.doi.org/10.1016/S0034-4257(96)00070-3]
108. Masoni, A.; Ercoli, L.; Mariotti, M. Spectral Properties of Leaves Deficient in Iron, Sulfur, Magnesium, and Manganese. Agron. J.; 1996; 88, pp. 937-943. [DOI: https://dx.doi.org/10.2134/agronj1996.00021962003600060015x]
109. Milton, N.M.; Eiswerth, B.A.; Ager, C.M. Effect of Phosphorus Deficiency on Spectral Reflectance and Morphology of Soybean Plants. Remote Sens. Environ.; 1991; 36, pp. 121-127. [DOI: https://dx.doi.org/10.1016/0034-4257(91)90034-4]
110. Motomiya, A.V.d.A.; Valente, I.M.Q.; Molin, J.P.; Motomiya, W.R.; Biscaro, G.A.; Jordan, R.A. Índice de Vegetação No Algodoeiro Sob Diferentes Doses de Nitrogênio e Regulador de Crescimento. Semin. Cienc. Agrar.; 2014; 35, pp. 169-178. [DOI: https://dx.doi.org/10.5433/1679-0359.2014v35n1p169]
111. Niaz, N.; Gulzar, S.; Kazmi, J.H.; Aleem, S.A.; Pham, M.-P.P.; Mierzwa-Hersztek, M.M.-H.; Riaz Ahmed, S.; Lahori, A.H.; Mushtaq, Z.N. Assessment of Chlorophyll Content in Leaves of Crops and Orchards Based on SPAD, Multispectral, and Hyperspectral Techniques. Ecol. Quest.; 2023; 35, pp. 161-174. [DOI: https://dx.doi.org/10.12775/EQ.2024.025]
112. Oliveira, R.A.; Näsi, R.; Niemeläinen, O.; Nyholm, L.; Alhonoja, K.; Kaivosoja, J.; Jauhiainen, L.; Viljanen, N.; Nezami, S.; Markelin, L.
113. Padilla, F.M.; Teresa Peña-Fleitas, M.; Gallardo, M.; Thompson, R.B. Evaluation of Optical Sensor Measurements of Canopy Reflectance and of Leaf Flavonols and Chlorophyll Contents to Assess Crop Nitrogen Status of Muskmelon. Eur. J. Agron.; 2014; 58, pp. 39-52. [DOI: https://dx.doi.org/10.1016/j.eja.2014.04.006]
114. Pandey, P.; Singh, S.; Khan, M.S.; Semwal, M. Non-Invasive Estimation of Foliar Nitrogen Concentration Using Spectral Characteristics of Menthol Mint (Mentha arvensis L.). Front. Plant Sci.; 2022; 13, 680282. [DOI: https://dx.doi.org/10.3389/fpls.2022.680282] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35615128]
115. Padilla, F.M.; Karaca, C.; Peña-Fleitas, M.T.; Gallardo, M.; Rodríguez, A.; Thompson, R.B. Cultivar Effect on Proximal Optical Sensor Measurements and Estimation of Leaf N Content in Muskmelon and Sweet Pepper. Eur. J. Agron.; 2024; 159, 127249. [DOI: https://dx.doi.org/10.1016/j.eja.2024.127249]
116. Putra, B.T.W. New Low-Cost Portable Sensing System Integrated with on-the-Go Fertilizer Application System for Plantation Crops. Measurement; 2020; 155, 107562. [DOI: https://dx.doi.org/10.1016/j.measurement.2020.107562]
117. Rodriguez, D.; Fitzgerald, G.J.; Belford, R.; Christensen, L.K. Detection of Nitrogen Deficiency in Wheat from Spectral Reflectance Indices and Basic Crop Eco-Physiological Concepts. Aust. J. Agric. Res.; 2006; 57, pp. 781-789. [DOI: https://dx.doi.org/10.1071/AR05361]
118. Roy Choudhury, M.; Christopher, J.; Das, S.; Apan, A.; Menzies, N.W.; Chapman, S.; Mellor, V.; Dang, Y.P. Detection of Calcium, Magnesium, and Chlorophyll Variations of Wheat Genotypes on Sodic Soils Using Hyperspectral Red Edge Parameters. Environ. Technol. Innov.; 2022; 27, 102469. [DOI: https://dx.doi.org/10.1016/j.eti.2022.102469]
119. Rustioni, L.; Grossi, D.; Brancadoro, L.; Failla, O. Characterization of Iron Deficiency Symptoms in Grapevine (Vitis spp.) Leaves by Reflectance Spectroscopy. Plant Physiol. Biochem.; 2017; 118, pp. 342-347. [DOI: https://dx.doi.org/10.1016/j.plaphy.2017.06.031] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28688348]
120. Sanaeifar, A.; Yang, C.; Min, A.; Jones, C.R.; Michaels, T.E.; Krueger, Q.J.; Barnes, R.; Velte, T.J. Noninvasive Early Detection of Nutrient Deficiencies in Greenhouse-Grown Industrial Hemp Using Hyperspectral Imaging. Remote Sens.; 2024; 16, 187. [DOI: https://dx.doi.org/10.3390/rs16010187]
121. Sharaf-Eldin, M.A.; Elsayed, S.; Elmetwalli, A.H.; Yaseen, Z.M.; Moghanm, F.S.; Elbagory, M.; El-Nahrawy, S.; Omara, A.E.-D.; Tyler, A.N.; Elsherbiny, O. Using Optimized Three-Band Spectral Indices and a Machine Learning Model to Assess Squash Characteristics under Moisture and Potassium Deficiency Stress. Horticulturae; 2023; 9, 79. [DOI: https://dx.doi.org/10.3390/horticulturae9010079]
122. Tang, Z.; Wang, X.; Xiang, Y.; Liang, J.; Guo, J.; Li, W.; Lu, J.; Du, R.; Li, Z.; Zhang, F. Application of Hyperspectral Technology for Leaf Function Monitoring and Nitrogen Nutrient Diagnosis in Soybean (Glycine max L.) Production Systems on the Loess Plateau of China. Eur. J. Agron.; 2024; 154, 127098. [DOI: https://dx.doi.org/10.1016/j.eja.2024.127098]
123. Wang, S.; Guan, K.; Wang, Z.; Ainsworth, E.A.; Zheng, T.; Townsend, P.A.; Liu, N.; Nafziger, E.; Masters, M.D.; Li, K.
124. Shen, Z.; Wang, K.; Wang, R. Feasibility of Field Evaluation of Rice Nitrogen Status from Reflectance Spectra of Canopy. Pedosphere A Q. J. Soil Sci.; 1998; 8, pp. 121-126.
125. Yi, J.; Krusenbaum, L.; Unger, P.; Hüging, H.; Seidel, S.J.; Schaaf, G.; Gall, J. Deep Learning for Non-Invasive Diagnosis of Nutrient Deficiencies in Sugar Beet Using RGB Images. Sensors; 2020; 20, 5893. [DOI: https://dx.doi.org/10.3390/s20205893]
126. Zhao, D.; Raja Reddy, K.; Kakani, V.G.; Read, J.J.; Carter, G.A. Corn (Zea mays L.) Growth, Leaf Pigment Concentration, Photosynthesis and Leaf Hyperspectral Reflectance Properties as Affected by Nitrogen Supply. Plant Soil; 2003; 257, pp. 205-218. [DOI: https://dx.doi.org/10.1023/A:1026233732507]
127. de Mello Prado, R. Mineral Nutrition of Tropical Plants; Springer International Publishing: Cham, Switzerland, 2021; 339p. ISBN 978-3-030-71261-7
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
The global transition to a digital era is crucial for society, as most daily activities are driven by digital technologies aimed at enhancing productivity and efficiency in the production of food, fibers, and bioenergy. However, the segregation of digital techniques and equipment in both rural and urban areas poses significant obstacles to technological efforts aimed at combating hunger, ensuring sustainable agriculture, and fostering innovations aligned with the United Nations Sustainable Development Goals (SDGs 02 and 09). Rural regions, which are often less connected to technological advancements, require digital transformation to shift from subsistence farming to market-integrated production. Recent efforts to expand digitalization in these areas have shown promising results. Digital agriculture encompasses terms such as artificial intelligence (AI), the Internet of Things (IoT), big data, and precision agriculture integrating information and communication with geospatial and satellite technologies to manage and visualize natural resources and agricultural production. This digitalization involves both internal and external property management through data analysis related to location, climate, phytosanitary status, and consumption. By utilizing sensors integrated into unmanned aerial vehicles (UAVs) and connected to mobile devices and machinery, farmers can monitor animals, soil, water, and plants, facilitating informed decision-making. An important limitation in studies on nutritional diagnostics is the lack of accuracy validation based on plant responses, particularly in terms of yield. This issue is observed even in conventional leaf tissue analysis methods. The absence of such validation raises concerns about the reliability of digital tools under real field conditions. To ensure the effectiveness of spectral reflectance-based diagnostics, it is essential to conduct additional studies in commercial fields across different regions. These studies are crucial to confirm the accuracy of these methods and to strengthen the development of digital and precision agriculture.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details
; Prado Renato de Mello 1
; Campos Cid Naudi Silva 2
; Baio Fábio Henrique Rojo 2
; Teodoro Larissa Pereira Ribeiro 2
; Teodoro, Paulo Eduardo 2
; Santana, Dthenifer Cordeiro 3
1 Department of Soil Science, São Paulo State University “Júlio de Mesquita Filho” UNESP/FCAV, Jaboticabal 14884-900, Brazil; [email protected]
2 Department of Agronomy, Federal University of Mato Grosso Do Sul (UFMS), Chapadão Do Sul 79560-000, Brazil; [email protected] (C.N.S.C.); [email protected] (F.H.R.B.); [email protected] (L.P.R.T.); [email protected] (P.E.T.)
3 Department of Agronomy, State University of São Paulo (UNESP), Ilha Solteira 15385-000, Brazil; [email protected]