Vinayaka1*
P. Rama Chandra Prasad2†
G. Avinash3†
Amaresh4†R. Arun Kumar5P. Murali1C. Palaniswami6P. Govindaraj4- 1Statistics and Economics Section, ICAR-Sugarcane Breeding Institute, Tamil Nadu, India
- 2Lab for Spatial Informatics, International Institute of Information Technology, Gachibowli, Telangana, India
- 3Executive Chairman & CEO, Avyagraha Research and Analytics LLP, Ramasagara, Karnataka, India
- 4Division of Crop Improvement, ICAR-Sugarcane Breeding Institute, Tamil Nadu, India
- 5Plant Physiology Section, ICAR-Sugarcane Breeding Institute, Tamil Nadu, India
- 6Soil Science Section, ICAR-Sugarcane Breeding Institute, Tamil Nadu, India
Sugarcane is a vital cash crop with substantial significance in both global sugar production and the biofuel industry. However, its sustainable cultivation faces persistent challenges from environmental stressors, particularly salinity and water scarcity. In recent years, the integration of artificial intelligence (AI) and remote sensing (RS) technologies has proven to be a transformative approach for detecting and evaluating these stress conditions, offering critical insights for advancing precision agriculture (PA). This review explores the utilization of satellite imagery and sensor-based data including RGB, multispectral, hyperspectral imaging, and unmanned aerial vehicles (UAVs) to monitor stress-related parameters in sugarcane farming. It emphasizes key indices used to assess water stress, generate salinity stress maps, and estimate nitrogen levels, demonstrating their role in equipping farmers with actionable information to optimize irrigation and nutrient management strategies. These innovations significantly enhance crop productivity while promoting environmental sustainability. The review sets out three core objectives: (i) to evaluate the contribution of AI and RS in assessing water stress in sugarcane cultivation, (ii) to examine methods for mapping salinity stress using RS and AI tools, and (iii) to highlight the relevance of spectral indices in tracking nitrogen status in sugarcane crops. Drawing upon reputable bibliographic sources such as Google Scholar, Scopus, ResearchGate, and Web of Science, along with current literature on AI and RS applications in sugarcane stress assessment, the review consolidates detailed information on advanced sensors and UAV technologies. It also introduces novel deep learning models and sensor platforms that have received limited attention in prior studies. In conclusion, the review affirms that AI-driven remote sensing is a highly effective approach for monitoring and managing critical stress factors in sugarcane production. It not only contributes to enhanced yield and crop quality but also delivers significant socio-economic and environmental benefits, marking a major step forward in achieving sustainable and efficient sugarcane cultivation.
1 Introduction
Sugarcane (Saccharum officinarum L.) belongs to the Andropogoneae species within the Poaceae family (Grof and Campbell, 2001). It is a crop of considerable economic importance, primarily cultivated in tropical and subtropical regions (Driemeier et al., 2016). According to recent estimates from the Food and Agriculture Organization (FAO) and the USDA, global raw sugar production for the 2023 24 marketing year is expected to surpass 183.5 million tonnes, marking a steady rise from 179 million tonnes in 2019 (FAO, 2022). Further projections estimate a production of approximately 186 million tonnes in 2024 25, spurred by increased output in countries like Thailand, India, and China (Infomerics Valuation and Rating Pvt. Ltd, 2024). Beyond its primary use in sugar production, sugarcane serves as a most important biological resource for creating ethanol, bagasse, fiber, molasses, rum, and cachaca (Vinayaka and Prasad, 2024; Amaresh et al., 2024; Suresha et al., 2024). Enhancing agricultural productivity remains a core focus of precision agriculture (PA), influenced by multiple environmental and agronomic factors (Amarasingam et al., 2022; Soltanikazemi et al., 2022). However, sugarcane cultivation is frequently challenged by abiotic stresses such as salinity and water scarcity, which negatively impact both yield and crop quality. The crop accounts for around 29% of global agricultural water use (Milagro Jorrat et al., 2018), and nearly 60% of sugarcane-growing areas in India experience water limitations often due to insufficient irrigation, canal closures during summer, and recurrent droughts (Sundara et al., 2002). Developing water-efficient sugarcane genotypes is essential for sustaining production while improving regional and global water security (Tayade et al., 2020; Kumar et al., 2020). Salinity poses a parallel threat: globally, approximately 33% of irrigated land and 20% of cultivated land are affected by saline conditions. In India, sugarcane is cultivated on nearly 5 million hectares, with about 25% of this area impacted by salinity, alkalinity, or saline irrigation water (Vasantha and Gomathi, 2012).
Remote sensing (RS) has become an indispensable tool in crop stress monitoring due to its ability to capture high-resolution, large-scale data (Huang et al., 2018). The accurate assessment of water, salinity, and nitrogen stress in sugarcane is decisive for informed crop management. Water deficits can significantly reduce productivity, underscoring the importance of precise irrigation scheduling (Hamzeh et al., 2013). Likewise, salinity stress caused by high salt concentrations in soil can hinder plant growth and yield potential (Hamzeh et al., 2016). Early detection allows for timely mitigation strategies such as leaching or improved drainage (Patil et al., 2021; Watanabe et al., 2022). Leaf nitrogen content (LNC) serves as a pivotal indicator of crop health and nutritional balance. Nitrogen deficiency limits growth and productivity, whereas excess application can cause environmental damage (Anas et al., 2020). Accurate monitoring of LNC supports optimized fertilization strategies, promoting sustainable practices (Virnodkar et al., 2020). Hence, evaluating water, salinity, and nitrogen stress is essential for improving sugarcane performance (Ferreira et al., 2017; Kumar et al., 2023b). Water stress in sugarcane is often attributed to erratic rainfall, exposure to alternating wet and dry periods, and high evapotranspiration rates (Bispo et al., 2022; Brunini and Turco, 2016). The concept of Crop Water Stress (CWS), introduced by Idso et al. (1981), has become a standard for assessing water deficits at both leaf and canopy levels. Salinity stress can arise from saline soil layers, the use of brackish irrigation water, elevated groundwater tables, or seawater intrusion (Hamzeh et al., 2016), contributing to widespread soil degradation in sugarcane-growing regions (Hamzeh et al., 2013). High water tables exacerbate this problem by raising salinity to damaging levels, thereby reducing crop output (Hamzeh et al., 2012). To mitigate such stresses, precision irrigation techniques and real-time monitoring of soil moisture and crop development are crucial (Yang H. et al., 2019). Additionally, salt removal from the root zone and field-specific corrective measures are necessary to maintain productivity (Burt and Isbell, 2005). Artificial intelligence (AI) and RS play complementary roles in assessing water and salinity stress in sugarcane. RS techniques such as thermal imaging and satellite-based data are valuable for estimating evapotranspiration (ET) and managing irrigation systems (Hamzeh et al., 2016; Das et al., 2020; Bispo et al., 2022; Watanabe et al., 2022). Energy balance models and vegetation indices are used to estimate ET and detect crop water stress (Patil et al., 2021; Virnodkar et al., 2021). Meanwhile, AI approaches, particularly machine learning (ML) and convolutional neural networks (CNNs), have shown promise in analyzing RS data to classify and map stress-affected zones in sugarcane fields. Nitrogen (N) remains a critical determinant of sugarcane yield and quality (Wiedenfeld, 1995), influencing attributes such as photosynthesis, tiller production, stem length, and girth (Gopalasundaram et al., 2012). Effective nitrogen management is thus central to sustainable productivity (Boschiero et al., 2020; Yang Y. et al., 2019). Extensive research on the optimal application of nitrogen fertilizers in sugarcane cultivation has been conducted globally, considering factors like growth cycle, climatic conditions, cultivar, and soil properties (Meyer et al., 1986; Wood et al., 1996). Although traditional techniques such as chlorophyll meters, soil sampling, and tissue analysis exist, they are often unsuitable for large-scale use due to their invasive, costly, and time-consuming nature (Ranjan et al., 2012).
Recent advances in AI and RS have facilitated non-invasive estimation of foliar nitrogen. Hyperspectral imagery, sensitive to biochemical changes in vegetation, has been effectively employed for nitrogen assessment (Soltanikazemi et al., 2022). Indices such as the Global Environmental Monitoring Index (GEMI), Chlorophyll Green Index (Clgreen), and Sentinel-2 Red-Edge Position Index (S2REP) have proven useful in estimating nitrogen levels in sugarcane leaves (Abdel-Rahman et al., 2013). Feature selection algorithms like Random Forest (RF) help reduce redundancy in hyperspectral datasets (Abdel-Rahman et al., 2010), while AI models including RF and Support Vector Regression (SVR) have demonstrated success in predicting nitrogen content rapidly and accurately (Martins et al., 2024). These methods offer practical, cost-effective, and scalable solutions for nitrogen monitoring (Bispo et al., 2022). Conventional stress monitoring techniques often lack the precision and speed required for proactive decision-making, reinforcing the need for advanced technologies. This review, therefore, emphasizes AI-based RS solutions for assessing key stressors in sugarcane agriculture. The primary objectives of this article are to: (i) explore how RS and AI are applied in sugarcane farming to evaluate crop water stress, (ii) investigate methods for mapping salinity stress using integrated RS and AI approaches, and (iii) highlight how spectral indices and AI tools can be leveraged to monitor the nitrogen status of sugarcane crops. Accordingly, the article is structured as follows: Section 2 covers bibliographic analysis; Section 3 presents a detailed discussion on integrated AI–RS methodologies and statistical software, along with recommendations for evaluating nitrogen, salinity, and water stress; Section 4 addresses existing challenges and limitations; Section 5 outlines future research directions; and Section 6 offers concluding insights.
2 Literary examination of sources
This study offers a comprehensive review of research focused on key stressors namely water stress, salinity stress, and plant nitrogen deficiency within the context of sugarcane cultivation. The review encompasses various aspects such as leaf nitrogen estimation, detection of water and salinity stress, and associated spectral signatures and vegetation indices. It particularly emphasizes the role of AI and RS technologies in addressing these challenges, highlighting recent advancements, practical applications, and the benefits and limitations of these approaches in sugarcane agriculture. To ensure a thorough evaluation, an extensive literature search was conducted using multiple academic databases, including Google Scholar, Scopus, ResearchGate, and Web of Science. The search focused on publications from 1981 to 2025, drawing upon peer-reviewed research articles, book chapters, and review papers. A total of 72 relevant sources (out of 88 references) were identified and systematically organized. Of these, 36 publications specifically demonstrated the effective application of AI, RS, or integrated approaches in various sugarcane-related domains such as water stress detection, leaf nitrogen monitoring, and salinity stress evaluation (Figure 1, Figure 2).
Figure 1. Year wise distribution of referenced studies downloaded from various databases. Field Crops Res.: Field Crops Research, Int. J Remote Sens.: International Journal of Remote Sensing, Int. J. Appl. Earth Obs. Geoinf.: International Journal of Applied Earth Observation and Geoinformation, Agric. Water Manag.: Agricultural Water Management, Symp or Proc or Conf: Symposium/Proceedings/Conference Papers, Comput Electron Agric.: Computers and Electronics in Agriculture, J Indian Soc Remote Sens: Journal of the Indian Society of Remote Sensing, Precis. Agric.: Precision Agriculture.
Figure 2. Number of referenced studies related to water, salinity, and nitrogen stress.
Journal titles have been abbreviated following the ISO4 standard, with a few exceptions such as Plan Plus, Water SA, Sustainability, Sugar Tech, and FAO Annual Report. The review also incorporates valuable content from symposium articles (Symp), conference proceedings (Proc), and conference papers (Conf). A strategic search methodology was adopted to compile the bibliography. Specific keywords and key phrases were used either individually or in logical combinations to locate relevant material. These included: “crop water stress”, “water stress of sugarcane assessment using AI/ML/DL and remote sensing”, “salinity stress”, “salinity stress of sugarcane assessment using AI/ML/DL and remote sensing”, “leaf nitrogen estimation in sugarcane using AI/ML/DL and remote sensing”, “drought detection”, “plant nitrogen stress”, “evapotranspiration”, “water productivity”, “water balance”, “water deficit”. These search terms enabled the retrieval of a broad and diverse set of studies on AI and RS applications in sugarcane management. Figure 3 illustrates a word frequency analysis of the selected references, focusing on the prominence of AI-based RS research in sugarcane stress management. Notably, several recent publications also served as entry points for accessing foundational works and earlier studies, offering deeper insight into the evolution of the field.
Figure 3. Keywords or phrases used for searching articles in the online platforms. GIS, Geographic Information System; RF, Random Forest; SVM, Support Vector Machine.
3 Detailed critiques
The utilization of RS applications in sugarcane farming encompasses a wide range of aspects, including crop classification, harvest planning, yield forecasting, disease detection and management (Palaniswami et al., 2011; 2014; Vinayaka and Prasad, 2024), assessment of crop health and growth, and detection of CWS. Among these, CWS detection plays a crucial role in predicting yield potential and optimizing irrigation scheduling across different growth stages and seasons. Various methodologies have been developed for identifying CWS, integrating soil water measurements, plant physiological responses, and RS techniques. The present study provides a comprehensive review of global approaches for detecting water stress in sugarcane using diverse RS methods and ML algorithms. The compiled indices (Table 1) illustrate the breadth of RS-based approaches employed for evaluating water status in irrigated sugarcane fields, reflecting the dynamic and adaptable nature of water stress assessment methods.
Table 1. Vegetation or spectral indices used in the referenced studies for crop water stress estimation.
Advancements in precision agriculture have further expanded the role of RS through the integration of AI and the Internet of Things (IoT), jointly termed AIoT-based water management. These systems leverage real-time data from soil moisture and weather sensors to optimize irrigation schedules, significantly enhancing water-use efficiency (Yueanket et al., 2024). Predictive algorithms such as Long Short-Term Memory (LSTM) models forecast irrigation requirements based on crop growth stages, thereby minimizing water wastage and promoting sustainable resource use. RS complements these systems by providing high-resolution imagery from satellites and drones for spatiotemporal monitoring of crop health and water stress through indices such as Normalized Difference Vegetation Index (NDVI) (Pawar et al., 2024). Moreover, thermal and hyperspectral imaging enable precise monitoring of environmental stressors, facilitating timely interventions (Swami et al., 2025; Cho et al., 2024).
Beyond water management, RS techniques have also been effectively employed for salinity stress and nitrogen status assessment in sugarcane. High soil salinity adversely affects plant growth, yield, and overall crop health, making its evaluation essential for sustainable production. Table 2 presents an overview of the RS-based indices and approaches commonly used for salinity stress assessment in sugarcane. Similarly, estimation of LNC is pivotal for optimizing fertilizer application and ensuring optimal growth and yield. Spectral reflectance measurements, vegetation indices, hyperspectral imaging, satellite and aerial imagery, and chlorophyll-based estimations have been widely applied for LNC evaluation (Table 3). AI-driven decision support systems further enhance these applications by integrating multi-source RS and IoT data to provide actionable insights for managing nitrogen levels and mitigating salinity impacts (Koohi et al., 2023; Cho et al., 2024). Despite these advancements, challenges such as data quality, accessibility for small-scale farmers, and economic feasibility persist, emphasizing the need for scalable and inclusive technological adoption in sugarcane agriculture.
Table 2. Vegetation or spectral indices used in the referenced studies for salinity stress assessment.
Table 3. Vegetation or spectral indices used in the referenced studies for nitrogen levels evaluation.
3.1 Water stress assessment
In the earlier studies, Rahman et al. (2004) employed NDVI to identify sugarcane areas and assess crop conditions, considering factors like leaf water content, nitrogen deficiency, pigments, foliar nutrients, and agronomic parameters. Abdel-Rahman and Ahmed (2008) found that the infrared/red ratio from Landsat TM NIR radiometer, SWIR bands, and the digital multispectral video (DMSV) sensor were effective in detecting water content in sugarcane crops. Detection of water stress, attributed to a reduction in the photosynthesis process, was found to be best achieved at the canopy level using VIS, red edge, and NIR regions (Berni et al., 2009). Brunini and Turco (2016) investigated sugarcane water stress indices in irrigated areas with different exposures and slopes. Their evaluation of daily water stress index and soil water potential revealed variations based on exposure and slope. The water stress index, derived from infrared thermometry, helped determine the optimal timing for irrigating sugarcane crops. Experiments conducted during various growing phases (tillering, growth, and maturation) on surfaces with slopes ranging from 0 to 40% and different solar exposures indicated that the ideal irrigation timing varied between 2.0 to 5.0 °C, depending on the sugarcane phase. A review by Katsoulas et al. (2016) focused on crop water stress and nutrient detection using reflectance measurement approaches and sensors in a greenhouse. They observed that ground-based sensor data indices were efficient for water stress detection but were influenced by factors such as leaf age, thickness, soil background, and canopy structure. Veysi et al. (2017) introduced a novel method for computing Crop Water Stress Index (CWSI) from satellite data, utilizing hot and cold pixels without the need for ground ancillary data. The study focused on irrigation scheduling in sugarcane during the growing season (May–September) and demonstrated superior performance compared to two alternative approaches, showing a strong coefficient of determination. The researchers observed a negative correlation between Vegetation Water Content (VWC) and CWSI, with R2 values ranging from 0.42 to 0.78. Validation of the new approach involved the analysis of eight Landsat 8 satellite images alongside ground truth data obtained through in situ measurements of canopy temperature and VWC.
An in-depth analysis of the studies in Table 4 reveals the progressive evolution of crop water stress assessment methodologies, integrating diverse data sources, modelling frameworks, and RS techniques. Early efforts, such as by Hellegers et al. (2009) and Singels et al. (2010), employed models like Surface Energy Balance Algorithm for Land (SEBAL) and CANEGRO for estimating ET and simulating physiological responses under stress, although they faced limitations in replicating yield-related processes. As studies progressed, RS technologies became central, with Teixeira et al. (2016) and Veysi et al. (2017, 2020) combining MODIS, Landsat, and SEBAL with meteorological and soil data, enabling improved spatial and temporal resolution in ET and water productivity estimates. Thermal-based indices, particularly the CWSI, emerged as powerful indicators in studies like Lebourgeois et al. (2010) and Farsi et al., demonstrating strong correlation with field-based measures of water deficit, despite challenges like cloud interference and calibration complexity. ML and deep learning (DL) significantly enhanced CWS assessment in later studies. For instance, Virnodkar et al. (2021) employed DenseResUNet and achieved high segmentation accuracy for stressed fields, while Alavi et al. (2024) and Melo et al. (2022) showed that advanced ML models like Random Forest and Inception-ResNet-v2 provided highly accurate predictions (R2 = 0.92–0.99, RMSE = 2.02–0.32 mmd−1) of crop water demand and thermal stress patterns. Notably, Gonçalves et al. (2022) and Bispo et al. (2022) showcased hybrid modeling approaches using Google Earth Engine - Surface Energy Balance Algorithm for Land (geeSEBAL) and Spatial Evapo-Transpiration Modeling Interface (SETMI), integrating RS with in-field micrometeorological data to refine ET estimates and irrigation management (RMSE = 0.46, R2 = 0.94 0.97). Across studies, spectral indices (e.g., TVDI: temperature vegetation dryness index; NDVI), soil moisture probes, thermal imaging, and energy balance models consistently contributed to assessing water stress, though limitations such as low-resolution meteorological data, cloud cover, and sensor calibration persisted. Collectively, these investigations underline a clear trend toward integrating multi-source RS data with AI/ML algorithms, enabling more precise, scalable, and real-time assessments of water stress in sugarcane agriculture supporting smarter irrigation scheduling and resilient crop management.
Table 4. Critical observations from referenced studies for crop water stress assessment.
3.2 Monitoring and estimation of salinity stress
Salinity is a critical factor affecting soil health and crop growth (Chele et al., 2021). Salinity stress assessment in sugarcane fields involves evaluating the effects of salinity on sugarcane growth, physiology, and yield. Various studies have been conducted to assess the impact of salinity stress on sugarcane. Kumar et al. (2023a) found that salinity stress significantly affected sugarcane yield, commercial cane sugar (CCS) yield, number of millable cane (NMC), single cane weight (SCW), and pol % in juice. Vu et al. (2023) demonstrated that the application of biochar had positive effects on the growth and physiology of sugarcane under both saline and non-saline conditions. Dhansu et al. (2022) conducted experiments to evaluate the response of popular sub-tropical sugarcane varieties to salinity stress and observed significant reductions in growth, relative water content (RWC), and gas exchange traits under saline conditions. Simoes et al. (2023) evaluated the growth-related traits of Saccharum genotypes under saline and non-saline conditions and identified promising genotypes with enhanced salinity tolerance. Djajadi et al. (2022) investigated the influence of salinity stress on sugarcane growth, soil nutrient content, and leaves and found that saline stress decreased soil organic and available K (Potassium), as well as the content of N and K in sugarcane leaves. Mohanan et al. (2021) also discussed the assessment of salinity stress tolerance in transgenic sugarcane plants overexpressing the Glyoxalase III gene. However, these studies not utilized the AI algorithms and RS data. As per the current advancements, AI algorithms applied to RS data can effectively identify and quantify soil salinity levels.
The combined analysis of studies in Table 5 demonstrates the growing effectiveness of integrating hyperspectral and multispectral RS data with ML algorithms to assess salinity stress in agricultural soils, particularly sugarcane fields. Hamzeh et al. (2012) laid foundational work by applying classifiers like Support Vector Machine (SVM), Spectral Angle Mapper (SAM), Minimum Distance (MD) and Maximum Likelihood algorithm (MLA) on Hyperion imagery in Khuzestan, Iran, where SVM outperformed others with a classification accuracy of 78.7%, revealing that RS-based salinity mapping can support targeted management of sugarcane in varied salinity zones. Building on this, Hamzeh et al. (2013) compared 21 hyperspectral vegetation indices (VIs), identifying optimized indices like OSAVI (Optimized soil-adjusted vegetation index) and VOG1 (Vogelmann red edge index) as reliable predictors of soil salinity with strong correlations (R² up to 0.69), underscoring the importance of selecting appropriate spectral bands sensitive to chlorophyll and water content. Hamzeh et al. (2016) further contrasted Hyperion and Landsat data, concluding that while Landsat delivered superior categorical classification accuracy (84.84%, Kappa 0.77), Hyperion was better suited for continuous salinity estimation. Notably, their findings introduced the Salinity and Water Stress Index (SWSI) as a robust metric for predicting salinity levels. Recent studies shifted toward advanced ML approaches for broader applicability. Haq et al. (2023) and Kaplan et al. (2023) demonstrated that ensemble learning methods such as RF Regression (RFR) and SVM using Landsat 8 and Sentinel-2 data respectively, achieved high accuracy in salinity prediction (R² = 0.94 and 0.84), confirming the scalability of RS-ML integration for local to regional salinity monitoring. Overall, the comparative findings highlight a consensus that hyperspectral data combined with tree-based ML models yield reliable predictions of salinity stress, with vegetation indices and spectral reflectance offering crucial insight. However, accuracy is often limited by sensor resolution, endmember variability, and the spatial distribution of ground truth data. Accurate yield predictions based on salinity levels can also help optimize resource allocation and improve overall crop management strategies, potentially increasing sugarcane yield and quality (Waters et al., 2025). These studies collectively highlight the value of tailored RS-ML frameworks in salinity stress management, enabling precision interventions in salt-affected sugarcane agroecosystems.
Table 5. Critical observations from referenced studies for salinity stress estimation.
3.3 Estimation of nitrogen levels
Reckoning leaf N using RS in sugarcane crops is important for optimizing nitrogen fertilizer management and improving crop growth and yield (Reyes-Trujillo et al., 2021). In recent years, different algorithms have been developed to combine spectral reflectance features with ML techniques for the estimation of leaf physiological parameters, including chlorophyll content, moisture content, and nitrogen content (Gai et al., 2023). RS techniques, such as digital image processing, hyperspectral data analysis, and drone multispectral imaging, can provide valuable information on leaf N concentrations and crop vigor. These techniques allow for non-destructive and fast estimation of nitrogen levels, enabling more precise and efficient fertilizer application (Lofton, 2012; Barros et al., 2021). By analyzing the correlation between LNC and various image features, regression models can be constructed to accurately estimate nitrogen content based on color and texture parameters (Li et al., 2022). Additionally, RS can help determine the optimum nitrogen rate and application timing for sugarcane production, ensuring that nitrogen is applied at the right time and in the right amount to maximize crop yield (Abebe et al., 2023).
The studies summarized in Table 6 emphasize the growing reliance on hyperspectral and multispectral RS combined with ML models for estimating nitrogen (N) content in sugarcane crops. Ahmed (2010) used hyperspectral RS in shade house trials to assess the impact of nitrogen and silicon treatments, identifying correlations between red-edge indices (e.g., R740/R720) and N levels. However, the study faced limitations such as weak reflectance-biochemical correlations and inconsistency between spectral and chemical sampling methods. Abdel-Rahman et al. (2010) advanced the analysis by using first-order derivative spectra to identify sensitive wavelengths, achieving an R² of 0.76 with the R743/R1316 ratio, yet noted the need for scalability to canopy-level applications. Expanding the scope, Abdel-Rahman et al. (2013) evaluated 163 hyperspectral bands using RF and stepwise multiple linear (SML) regression models, both yielding reliable predictions (RF: R² = 0.67, RMSEV = 0.15%; SML: R² = 0.71, RMSEV = 0.19%), demonstrating the potential of hyperspectral data for accurate N monitoring. In a more recent study, Soltanikazemi et al. (2022) utilized Sentinel-2 imagery and calculated multiple vegetation indices (e.g., S2REP, IRECI: Inverted Red-Edge Chlorophyll Index, NDVI) using RF and SVR, achieving modest performance (R² = 0.59, RMSE = 0.08), with RF slightly outperforming SVR (R² = 0.58, RMSE = 0.09). They emphasized the benefit of larger ground datasets and multi-temporal imagery to improve robustness. Finally, Martins et al. (2024) provided a detailed comparison of vegetation indices such as BNi: Buschman and Nagel index, NDRE: Normalized Difference Red-Edge, GNDVI: Green NDVI, and RI-1db: Ratio Index, all yielding R² > 0.65 and RMSE< 3.7 g/kg. However, environmental variability across locations and seasons influenced the predictive capacity of their models, with performance dropping by up to 50% for the same variety (SP 81 3250). Together, these studies indicate that while RS-based N estimation in sugarcane is promising, particularly with indices targeting red-edge and near-infrared regions, the accuracy is influenced by factors like sensor resolution, environmental heterogeneity, sample consistency, and model type. The integration of ML, especially ensemble models like RF, enhances predictive reliability, but operational deployment still requires more stable, scalable, and temporally adaptive frameworks for practical field use.
Table 6. Critical observations from referenced studies for nitrogen content evaluation.
3.4 Software application
This review systematically examined all the referenced studies to identify the use of specific statistical software, programming languages, or analytical tools for implementing AI-RS methodologies related to the estimation of water stress, salinity stress, and LNC in sugarcane farming. Six studies explicitly mentioned the use of software or coding frameworks. These include SYSTAT (Inman-Bamber, 2004), QGIS Desktop (Triadi et al., 2020), R, Python, and MATLAB (Virnodkar et al., 2020; 2021), R and MATLAB (Watanabe et al., 2022), Microsoft Excel 2019 and Python’s Scikit-learn library (Gai et al., 2023), and Environment for Visualizing Imagery (ENVI) 5.3.1 for image visualization and analysis (Alavi et al., 2024). Although many studies reported the application of RS and ML methods, only a few provided explicit details of the software environments or coding frameworks used.
3.5 Recommendations: best solutions from referenced studies
A critical synthesis of previous research (1981–2025) highlights several methodological approaches that stand out as best-practice solutions for stress detection in sugarcane agriculture. Thermal and hyperspectral remote sensing combined with energy balance models, such as SEBAL and CWSI, have provided reliable estimates of evapotranspiration and crop water stress, especially when integrated with meteorological and soil data to support irrigation scheduling (Hellegers et al., 2009; Veysi et al., 2017; Teixeira et al., 2016; Gonçalves et al., 2022). Recent advances demonstrate that deep learning models, including DenseResUNet and Inception-ResNet-v2, applied to high-resolution Sentinel-2 and UAV imagery, significantly improve canopy-level water stress segmentation and thermal pattern prediction (Virnodkar et al., 2021; Alavi et al., 2024; Melo et al., 2022). For salinity stress assessment, the combination of hyperspectral imagery with SVM classifiers and tree-based ensemble methods such as RFR has consistently yielded high classification accuracy and predictive performance across diverse environments (Hamzeh et al., 2012, 2013; Haq et al., 2023; Kaplan et al., 2023). Similarly, for nitrogen estimation, red-edge and NIR vegetation indices integrated with RF or SVR have shown strong correlations with leaf nitrogen content, offering scalable and non-destructive nutrient monitoring solutions (Abdel-Rahman et al., 2010, 2013; Soltanikazemi et al., 2022; Martins et al., 2024). In addition, data fusion approaches and cloud-based platforms like Google Earth Engine have enabled the integration of multi-sensor datasets, improving spatiotemporal resolution and analytical efficiency (Gonçalves et al., 2022; Bispo et al., 2022). Jointly, these approaches represent the most effective, validated solutions for operationalizing AI-RS frameworks in precision water, salinity, and nutrient management for sugarcane.
4 Challenges and limitations
Despite significant advancements, several challenges and limitations persist in the application of RS and AI technologies for assessing water stress, salinity stress, and LNC in sugarcane crops. These challenges primarily arise from the complexity of environmental conditions, limitations in sensor technology, data processing requirements, and the need for robust model calibration and validation.
4.1 Technical and sensor limitations
Spatial resolution remains a major constraint, as RS instruments often struggle to capture detailed information at the individual plant or plot level in large sugarcane plantations. Similarly, limited temporal resolution hampers the ability to monitor short-term fluctuations in water stress, salinity, and nutrient dynamics, which are critical for understanding crop responses across growth stages. While hyperspectral sensors can provide improved spectral resolution, they often face limitations in spatial coverage, data volume, and operational costs. Low radiometric sensitivity (radiometric resolution) can also limit the detection of subtle crop stress signals, especially in early stages. Cloud cover further complicates consistent monitoring by obstructing satellite imagery, particularly in regions prone to frequent cloudiness. In addition, high-quality RS data from advanced sensors can be expensive and less accessible, restricting adoption among resource-limited farmers and researchers.
4.2 Environmental and field variability
Variations in soil type, crop age, plant density, and microclimatic conditions significantly influence spectral reflectance, complicating the accurate discrimination of stress signals (Waters et al., 2025). Non-crop interference, such as background soil reflectance or surrounding vegetation, may introduce noise that obscures the true radiometric signals from sugarcane canopies (Som-Ard et al., 2021). Such variability makes it challenging to develop generalized models that can accurately capture stress patterns across diverse environments.
4.3 Data processing and integration challenges
The processing of RS data involves extensive pre-processing steps such as atmospheric correction, radiometric calibration, and cloud masking to ensure accuracy, which can be time-consuming and computationally demanding (Som-Ard et al., 2021). Integrating data from multiple sensors or platforms for comprehensive stress assessment requires sophisticated algorithms and standardized analysis-ready data (ARD) formats, which are still evolving. Obtaining reliable ground truth data for model calibration and validation remains another challenge, especially across large or remote agricultural regions.
4.4 Modelling and analytical constraints
ML and deep learning models, though promising, face constraints when applied to heterogeneous agricultural datasets. Unbalanced data distributions, intra-species variability, and insufficient ground truth samples can reduce model accuracy and generalizability (Kamarudin et al., 2021; Koohi et al., 2023). Effective calibration and validation are essential for improving model reliability across different environmental and management conditions.
Addressing these limitations requires continuous research, technological innovation, and interdisciplinary collaboration among RS scientists, agronomists, data analysts, and policymakers. Ongoing integration of AI and RS approaches holds great potential to overcome current barriers and enhance the precision and scalability of stress detection in sugarcane farming. The present review also emphasizes previous research that combines these emerging technologies to evaluate crop stresses, specifically water, salinity, and nitrogen in sugarcane cultivation systems.
5 Future directions
The future of sustainable agriculture depends on the widespread integration of advanced technologies such as ML, RS, IoT, robotics, PA, and cloud computing. Although these technologies possess immense potential, their adoption particularly in developing regions remains limited. In the context of increasing challenges such as climate change, land degradation, and water scarcity, the deployment of intelligent systems is essential to ensure efficient management of water and nutrient resources during crop production. Real-time monitoring and predictive analytics play a pivotal role in addressing critical stressors such as water deficiency, salinity, and nitrogen imbalance. When applied to RS data, ML techniques can effectively detect spatial variability in water use efficiency, particularly in low-productivity zones, thereby improving irrigation scheduling and enhancing crop management without expanding cultivated land or increasing water consumption. Integration of advanced RS technologies has further revolutionized sugarcane monitoring and management. Hyperspectral imaging enables precise detection of water and nutrient levels, improving the ability to assess crop health and stress conditions (Swami et al., 2025). The use of unmanned aerial vehicles (UAVs) and CubeSats provides high-resolution spatial and temporal data crucial for real-time monitoring of environmental stressors (Swami et al., 2025). Data fusion techniques, which combine information from multiple sensors and platforms, enhance the accuracy of assessing crop conditions and resource availability (Swami et al., 2025). AI-driven predictive modelling further strengthens decision-making in precision agriculture. The ML algorithms enhance the prediction of crop responses to water and nutrient availability, supporting site-specific management strategies (Gupta et al., 2024). Real-time monitoring systems, such as SWARM, dynamically adjust irrigation and nutrient delivery based on live data, optimizing resource use and improving efficiency (Babu et al., 2006).
Moreover, coupling RS datasets with ML classification algorithms especially those utilizing full spectral ranges has shown high accuracy in assessing salinity and nitrogen levels in sugarcane crops. Integrating explainable AI (XAI) approaches is crucial to enhance model transparency and interpretability. XAI provides insights into the decision-making process of ML models, allowing agronomists and farmers to understand which spectral features or environmental factors drive specific predictions. This interpretability builds trust in AI-driven recommendations and promotes informed field management decisions. However, several barriers remain. The high cost of proprietary cognitive farming solutions, coupled with the computational demands of advanced ML algorithms, continues to limit large-scale adoption. To overcome these challenges, the development of open-source, farmer-friendly platforms is essential to democratize access to digital technologies and reduce dependency on expensive commercial software. Furthermore, optimizing ML models to reduce processing time and computational overhead through lightweight algorithms and edge computing will enable integration with RS systems on resource-constrained devices. Future research should also prioritize sensor calibration, data accessibility, and ethical data management, ensuring privacy and equitable use of agricultural data. Establishing clear frameworks for data ownership and protection will foster trust and encourage the broader adoption of AI- and RS-based technologies. Collectively, these innovations will pave the way toward scalable, cost-effective, and sustainable precision management systems for water, salinity, and nutrient optimization in sugarcane cultivation (Figure 4).
Figure 4. Blueprint for future research: A flowchart for precisely estimating crop stress factors through AI-driven remote sensing. LiDER, Light Detection and Ranging; UAVs, Unmanned Aerial Vehicles; VIF, Variance Inflation Factor; ML, Machine Learning; DL, Deep Learning; RF, Random Forest; DCNN, Deep Convolutional Neural Networks.
6 Concluding remarks
Over the past decade, agricultural systems have increasingly integrated AI and RS technologies to address critical challenges and boost productivity. This review compiles current knowledge on the application of AI and RS in sugarcane cultivation, highlighting their strengths and constraints in managing crop water stress, leaf nitrogen estimation, and salinity stress mapping. Although their adoption in agriculture lags behind other sectors, the continuous evolution of sensors, UAVs, and ML algorithms presents substantial opportunities for innovation in sugarcane farming. Despite these advancements, significant barriers remain including high initial investment, sensor limitations, complex data processing requirements, limited technical expertise, and farmer apprehension. Nevertheless, AI-driven RS approaches show great promise for improving both yield and quality in sugarcane, while contributing positively to environmental and socio-economic outcomes.
Conventional soil moisture monitoring methods are often constrained by high sensor costs, installation complexity, and inaccuracies, especially across varied soil types and crop systems. Plant-based assessments, while more reliable and accurate, often lack scalability and are time-intensive. Research consistently shows that remotely sensed indices such as the Photochemical Reflectance Index (PRI) and NDVI are significantly correlated with physiological parameters like leaf water potential (LWP), stomatal conductance, crop coefficient, and stem water potential. However, relying solely on single-parameter metrics for plant water status estimation often falls short in precision. Among water stress indicators, the CWSI, particularly when derived from Earth Observation (EO) data has emerged as a preferred metric for assessing sugarcane water status across local to regional scales. Infrared thermometers continue to serve as effective ground tools for measuring canopy temperature, providing critical validation for RS-based observations. Midday stem water potential also remains a standard reference point for cross-verifying RS-derived estimates. For salinity stress evaluation, studies endorse the use of multispectral (e.g., Landsat ETM+) and hyperspectral (e.g., Hyperion) satellite data, with classification algorithms like minimum distance (MD) consistently delivering reliable results. Nitrogen estimation in sugarcane has similarly benefited from vegetation indices that combine near-infrared, green, and red-edge wavelengths. Indices such as SAVI (soil adjusted vegetation index), MSAVI (modified SAVI), NDVI, and OSAVI have been effectively incorporated into linear, nonlinear, and ML-based models, with ML approaches consistently outperforming traditional methods in accuracy and adaptability.
Among ML techniques, Artificial Neural Networks (ANNs) have shown particular success in analyzing UAV multispectral data and in determining variable contributions to target outcomes. Support Vector Machines (SVM) and Random Forests (RF) also demonstrate high potential in RS-based classification, though their full capabilities in detecting crop stresses specifically water, salinity, and nitrogen are still underexplored. Recent innovations like oblique and rotation-based RF classifiers have exhibited improved performance across varied datasets. The oblique RF approach, effective with discrete factorial features, is promising for evaluating water stress and warrants deeper investigation. Similarly, the rotation RF model, which integrates multiple rotated feature spaces, has surpassed traditional classifiers such as RF, SVM, and k-NN in several studies. Despite this, both rotation RF and deep CNNs remain underutilized in stress assessment involving RS data sources such as microwave imagery, UAVs, and Light Detection and Ranging (LiDAR). Machine learning’s capacity to aggregate and analyze data from diverse sources including ground measurements, sensor networks, meteorological data, and RS platforms like satellites, drones, and airborne systems makes it central to the future of digital agriculture. While ML has been widely applied in tasks such as crop classification, yield prediction, and condition monitoring, more targeted research is needed to fully harness its potential for stress assessments in sugarcane, particularly in water, salinity, and nitrogen dynamics. These areas are vital for informed irrigation planning and sustainable crop management, and thus demand greater attention from the research community. Continued advancements in this field will significantly benefit sugarcane agriculture by enhancing productivity, supporting long-term sustainability, and improving resilience against challenges such as climate variability, resource constraints, and market instability.
Author contributions
V: Conceptualization, Methodology, Data curation, Formal Analysis, Writing – original draft, Writing – review & editing. PP: Conceptualization, Data curation, Supervision, Writing – original draft, Writing – review & editing, Formal Analysis, Methodology. GA: Data curation, Formal Analysis, Methodology, Writing – review & editing, Writing – original draft. A: Data curation, Methodology, Writing – original draft, Writing – review & editing. RA: Conceptualization, Writing – original draft, Writing – review & editing. PM: Conceptualization, Writing – original draft, Writing – review & editing. CP: Conceptualization, Writing – original draft, Writing – review & editing. PG: Supervision, Writing – original draft, Writing – review & editing.
Funding
The author(s) declare that no financial support was received for the research, and/or publication of this article.
Acknowledgments
Research was supported by the Indian Council of Agricultural Research, Department of Agricultural Research and Education, Government of India. The authors thank the editors and reviewers for the critical suggestions that led improvements in the article.
Conflict of interest
Authors GA was employed by Avyagraha Research and Analytics LLP.
The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The handling editor SR declared a past co-authorship with the author(s) GA.
Generative AI statement
The author(s) declare that no Generative AI was used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
Abdel-Rahman E. M., Ahmed F. B. (2008). The application of remote sensing techniques to sugarcane (Saccharum spp. hybrid) production: A review of the literature. Int. J. Remote Sens. 29, 3753–3767. doi: 10.1080/01431160701874603
Abdel-Rahman E. M., Ahmed F. B., Berg. M. (2010). Estimation of sugarcane leaf nitrogen concentration using in situ spectroscopy. Int. J. Appl. Earth Observation Geoinformation 12, 52–57. doi: 10.1016/j.jag.2009.11.003
Abdel-Rahman E. M., Ahmed F. B., Ismail. R. (2013). Random forest regression and spectral band selection for estimating sugarcane leaf nitrogen concentration using EO-1 Hyperion hyperspectral data. Int. J. Remote Sens. 34, 712–728. doi: 10.1080/01431161.2012.713142
Abebe G., Tadesse T., Gessesse. B. (2023). Estimating leaf area index and biomass of sugarcane based on Gaussian process regression using Landsat 8 and Sentinel 1A observations. Int. J. Image Data Fusion 14, 58–88. doi: 10.1080/19479832.2022.2055157
Ahmed F. B. (2010). Estimation of leaf nitrogen and silicon using hyperspectral remote sensing. J. Appl. Remote Sens. 4, 043560. doi: 10.1117/1.3525241
Alavi M., Albaji M., Golabi M., Zareian M., Ayoubi S., Javanmard. R. (2024). Estimation of sugarcane evapotranspiration from remote sensing and limited meteorological variables using machine learning models. J. Hydrology 629, 130605. doi: 10.1016/j.jhydrol.2023.130605
Amarasingam N., Ashan Salgadoe A. S., Powell K., Gonzalez L. F., Natarajan. S. (2022). A review of UAV platforms, sensors, and applications for monitoring of sugarcane crops. Remote Sens. Applications: Soc. Environ. 26, 100712. doi: 10.1016/j.rsase.2022.100712
Amaresh N., Vinayaka A., Kumari S., Avinash G., Abhishek G. J., Dhanya V. G., et al. (2024). “Sugarcane Bagasse: Transforming Waste into Renewable and Bio-Compostable Materials for a Sustainable Future,” in Value Addition and Product Diversification in Sugarcane. Eds. Suresha G. S., Krishnappa G., Palanichamy M., Mahadeva Swamy H. K., Kuppusamy H., Govindakurup H. (Springer, Singapore). doi: 10.1007/978-981-97-7228-5_12
Anas M., Liao F., Verma K. K., Sarwar M. A., Mahmood A., Chen Z. L., et al. (2020). Fate of nitrogen in agriculture and environment: Agronomic, eco-physiological and molecular approaches to improve nitrogen use efficiency. Biol. Res. 53, 47. doi: 10.1186/s40659-020-00312-4
Babu C., Athmaraman N., Ganesan A., Soundararajan S., Muthukumar R. (2006). “SWARM: Sensor driven water and agricultural resources management,” in 2006 Annual IEEE India Conference. New Delhi, India:IEEE. 1–6. doi: 10.1109/INDCON.2006.302766
Barros P. P. D. S., Fiorio P. R., Demattê J. A. D. M., Martins J. A., Montezano Z. F., Dias. F. L. F. (2021). Estimation of leaf nitrogen levels in sugarcane using hyperspectral models. Ciec. Rural 52. doi: 10.1590/0103-8478cr20200630
Berni J. A., Zarco-Tejada P. J., Suárez L., Sepulcre-Cantó F., Morales D. I., Sobrino J. M., et al. (2009). Thermal and narrowband multispectral remote sensing for vegetation monitoring from an unmanned aerial vehicle. IEEE Trans. Geosci. Remote Sens. 47, 722–738. doi: 10.1109/TGRS.2008.2010457
Bispo R. C., Hernandez F. B. T., Gonçalves I. Z., Neale C. M. U., Teixeira. A. H. C. (2022). Remote sensing-based evapotranspiration modeling for sugarcane in Brazil using a hybrid approach. Agric. Water Manage. 271, 107763. doi: 10.1016/j.agwat.2022.107763
Boschiero B. N., Mariano E., Torres-Dorante L. O., Sattolo T. M., Otto R., Garcia P. L., et al. (2020). Nitrogen fertilizer effects on sugarcane growth, nutritional status, and productivity in tropical acid soils. Nutrient Cycling 117, 367–382. doi: 10.1007/s10705-020-10074-w
Brunini R. G., Turco J. E. (2016). Water stress indices for the sugarcane crop on different irrigated surfaces. Rev. Bras. Engenharia Agrícola e Ambiental 20, 925–929. doi: 10.1590/1807-1929/agriambi.v20n10p925-929
Burt C. M., Isbell B. (2005). Leaching of accumulated soil salinity under drip irrigation. Trans. ASAE 48, 2115–2121. doi: 10.13031/2013.20097
Chele K. H., Tinte M. M., Piater L. A., Dubery I. A., Tugizimana. F. (2021). Soil salinity, a serious environmental issue and plant responses: A metabolomics perspective. Metabolites 11, 724. doi: 10.3390/metabo11110724
Cho S. B., Soleh H. M., Choi J. W., Hwang W. H., Lee H., Cho Y. S., et al. (2024). Recent methods for evaluating crop water stress using ai techniques: A review. Sensors 24, 6313. doi: 10.3390/s24196313
Das K., Khiripet N., Chattanrassamee P., Kijkullert C., Veerachit V. (2020). Crop evapotranspiration estimates for sugarcane based on remote sensing and land surface model in Thailand. 2020 IEEE Int. Geosci. Remote Sens. Symposium (IGARSS), 5175–5178. doi: 10.1109/IGARSS39084.2020
Dhansu P., Kumar R., Kumar A., Vengavasi K., Raja A. K., Vasantha S., et al. (2022). Differential physiological traits, ion homeostasis, and cane yield of sub-tropical sugarcane varieties in response to long-term salinity stress. Sustainability 14, 13246. doi: 10.3390/su142013246
Djajadi D., Syaputra R., Hidayati. S. N. (2022). “Effect of salt stress on nutrient content in soil and leaves of three sugarcane varieties,” in IOP Conference Series: Earth and Environmental Science, Vol. 974. 012048. Bogor, Indonesia: IOP Publishing. doi: 10.1088/1755-1315/974/1/012048
Driemeier C., Ling L. Y., Sanches G. M., Pontes A. O., Magalhães P. S. G., Ferreira. J. E. (2016). A computational environment to support research in sugarcane agriculture. Comput. Electron. Agric. 130, 13–19. doi: 10.1016/j.compag.2016.10.002
FAO (2022). World food and agriculture – Statistical yearbook 2022 (FAO) Rome, Italy. doi: 10.4060/cc2211en
Ferreira T. H. S., Tsunada M. S., Bassi D., Araújo P., Mattiello L., Guidelli G. V., et al. (2017). Sugarcane water stress tolerance mechanisms and its implications on developing biotechnology solutions. Front. Plant Sci 8. doi: 10.3389/fpls.2017.01077
Gai J., Wang J., Xie S., Xiang L., Wang Z.Spectroscopic determination of chlorophyll content in sugarcane leaves for drought stress detectionPrecision Agriculture (2024)25:543–569.doi: 10.1007/s11119-023-10082-0
Gonçalves I. Z., Ruhoff A., Laipelt L., Bispo R. C., Hernandez F. B. T., Neale C. M. U., et al. (2022). Remote sensing-based evapotranspiration modeling using geeSEBAL for sugarcane irrigation management in Brazil. Agric. Water Manage. 274, 107965. doi: 10.1016/j.agwat.2022.107965
Gopalasundaram P., Bhaskaran A., Rakkiyappan. P. (2012). Integrated nutrient management in sugarcane. Sugar Tech 14, 3–20. doi: 10.1007/s12355-011-0097-x
Grof C. P. L., Campbell J. A. (2001). Sugarcane sucrose metabolism: Scope for molecular manipulation. Funct. Plant Biol. 28, 1. doi: 10.1071/PP00039
Gupta D. K., Pagani A., Zamboni P., Singh A. K. (2024). AI-powered revolution in plant sciences: advancements, applications, and challenges for sustainable agriculture and food security. Explor. Foods Foodomics 2, 443–459. doi: 10.37349/eff.2024.00045
Hamzeh S., Naseri A. A., AlaviPanah S. K., Bartholomeus H., Herold M. (2016). Assessing the accuracy of hyperspectral and multispectral satellite imagery for categorical and quantitative mapping of salinity stress in sugarcane fields. Int. J. Appl. Earth Observation Geoinformation 52, 412–421. doi: 10.1016/j.jag.2016.06.024
Hamzeh S., Naseri A. A., AlaviPanah S. K., Mojaradi B., Bartholomeus H. M., Clevers J. G. P. W., et al. (2013). Estimating salinity stress in sugarcane fields with spaceborne hyperspectral vegetation indices. Int. J. Appl. Earth Observation Geoinformation 21, 282–290. doi: 10.1016/j.jag.2012.07.002
Hamzeh S., Naseri A. A., Panah S. A., Mojaradi B., Bartholomeus H. M., Herold. M. (2012). Mapping salinity stress in sugarcane fields with hyperspectral satellite imagery. Remote Sens. Agriculture Ecosystems Hydrology XIV 8531, 528–535. doi: 10.1117/12.981655
Haq Y. U., Shahbaz M., Asif H. S., Al-Laith A., Alsabban W. H. (2023). Spatial mapping of soil salinity using machine learning and remote sensing in Kot Addu, Pakistan. Sustainability 15, 12943. doi: 10.3390/su151712943
Hellegers P. J. G. J., Soppe R., Perry C. J., Bastiaanssen. J. (2009). Combining remote sensing and economic analysis to support decisions that affect water productivity. Irrigation Sci 27, 243–251. doi: 10.1007/s00271-008-0139-7
Huang Y., Chen Z. X., Tao Y. U., Huang X. Z., Gu. X. F. (2018). Agricultural remote sensing big data: Management and applications. J. Integr. Agric. 17, 1915–1931. doi: 10.1016/S2095-3119(17)61859-8
Idso S. B., Jackson R. D., Pinter P. J., Reginato R. J., Hatfield. J. L. (1981). Normalizing the stress-degree-day parameter for environmental variability. Agric. Meteorology 24, 45–55. doi: 10.1016/0002-1571(81)90032-7
Infomerics Valuation and Rating Pvt. Ltd (2024). Sugar Industry Outlook – December. Available online at: https://www.infomerics.com/admin/uploads/sugar-industry-dec24.pdf. (Accessed May 25, 2025).
Inman-Bamber N. G. (2004). Sugarcane water stress criteria for irrigation and drying off. Field Crops Res. 89, 107–122. doi: 10.1016/j.fcr.2004.01.018
Kamarudin M. H., Ismail Z. H., Saidi N. B. (2021). Deep learning sensor fusion in plant water stress assessment: A comprehensive review. Appl. Sci. 11, 1403. doi: 10.3390/app11041403
Kaplan G., Gašparović M., Alqasemi A. S., Aldhaheri A., Abuelgasim A., Ibrahim M. (2023). Soil salinity prediction using machine learning and sentinel–2 remote sensing data in hyper–arid areas. Phys. Chem. Earth 130, 103400. doi: 10.1016/j.pce.2023.103400
Katsoulas N., Elvanidi A., Ferentinos K. P., Kacira G., Bartzanas C., Kittas. C. K. (2016). Crop reflectance monitoring as a tool for water stress detection in greenhouses: A review. Biosyst. Eng. 151, 374–398. doi: 10.1016/j.biosystemseng.2016.10.003
Koohi E., Gumiere S. J., Bonakdari H. (2023). “Artificial Intelligence Models for Detecting Spatiotemporal Crop Water Stress in schedule Irrigation: A review,” in EGU General Assembly Conference Abstracts. Vienna, Austria. EGU–3997. doi: 10.5194/egusphere-egu23-3997
Kumar R., Dhansu P., Kulshreshtha N., Meena M. R., Kumaraswamy M. H., Appunu C., et al. (2023a). Identification of salinity-tolerant stable sugarcane cultivars using AMMI, GGE and some other stability parameters under multi-environments of salinity stress. Sustainability 15, 1119. doi: 10.3390/su15021119
Kumar R., Sagar V., Verma V. C., Kumari M., Gujjar R. S., Goswami S. K., et al. (2023b). Drought and salinity stresses induced physio-biochemical changes in sugarcane: An overview of tolerance mechanism and mitigating approaches. Front. Plant Sci 14. doi: 10.3389/fpls.2023.1225234
Kumar R., Vasantha S., Tayade A., Anusha S., Geetha P., Hemaprabha. G. (2020). Physiological efficiency of sugarcane clones under water-limited conditions. Trans. ASABE 63, 133–140. doi: 10.13031/trans.13550
Lebourgeois V., Chopart J. L., Bégue A., Clouvel R., Courties J., Wery. P. (2010). Towards using a thermal infrared index combined with water balance modelling to monitor sugarcane irrigation in a tropical environment. Agric. Water Manage. 97, 75–82. doi: 10.1016/j.agwat.2009.08.013
Li X., Ba Y., Zhang M., Nong M., Yang C., Zhang. S. (2022). Sugarcane nitrogen concentration and irrigation level prediction based on UAV multispectral imagery. Sensors 22, 2711. doi: 10.3390/s22072711
Lofton J. (2012). Improving nitrogen management in sugarcane production of the Mid-South using remote sensing technology. Louisiana State University and Agricultural and Mechanical College, Baton Rouge, Louisiana, United States.
Martins J. A., Fiorio P. R., Silva C. A. A. C., Demattê J. A. M., Silva Barros. P. P. D. (2024). Application of vegetative indices for leaf nitrogen estimation in sugarcane using hyperspectral data. Sugar Tech 26, 160–170. doi: 10.1007/s12355-023-01329-1
Melo L. L. D., Melo V.G.M.L.D., Marques P. A. A., Frizzone J. A., Coelho R. D., Romero R. A. F., et al. (2022). Deep learning for identification of water deficits in sugarcane based on thermal images. Agric. Water Manage. 272, 107820. doi: 10.1016/j.agwat.2022.107820
Meyer J. H., Wood R. A., Leibbrandt. N. B. (1986). Recent advances in determining the N requirement of sugarcane in the South African sugar industry. Proc. South Afr. Sugar Technologists’ Assoc. 60, 205–211.
Milagro Jorrat M., Araujo P. Z., Mele. F. D. (2018). Sugarcane water footprint in the province of Tucumán, Argentina. Comparison between different management practices. J. Cleaner Production 188, 521–529. doi: 10.1016/j.jclepro.2018.03.242
Mohanan M. V., Pushpanathan A., Padmanabhan S., Sasikumar T., Jayanarayanan A. N., Selvarajan D., et al. (2021). Overexpression of Glyoxalase III gene in transgenic sugarcane confers enhanced performance under salinity stress. J. Plant Res. 134, 1083–1094. doi: 10.1007/s10265-021-01300-9
Mokhele T. A., Ahmed F. B. (2010). Estimation of leaf nitrogen and silicon using hyperspectral remote sensing. J. Appl. Remote Sens. 4 (1), 043560. doi: 10.1117/1.3525241
Palaniswami C., Gopalasundaram P., Bhaskaran. A. (2011). Application of GPS and GIS in sugarcane agriculture. Sugar Tech 13, 360–365. doi: 10.1007/s12355-011-0098-9
Palaniswami C., Viswanathan R., Bhaskaran A., Rakkiyappan P., Gopalasundaram. P. (2014). Mapping sugarcane yellow leaf disease affected area using remote sensing technique. J. Sugarcane Res. 4, 55–61.
Patil V. C., Jha S. K., Banerjee A., Algat. Y. (2021). “Sugarcane crop water status mapping using Landsat 8 imagery, crop water stress index (CWSI), and metric models,” in International Conference on Sugarcane Research: Sugarcane for Sugar and Beyond (CaneCon-2021), Coimbatore, India.
Pawar M. M., Pawar M. M., Pawar P. M., Deshmukh B. M. (2024). Advancing sugarcane farm management through NDVI-based color mapping and drone imaging. Logistic Operation Manage. Res. (LOMR) 3, 55–68. doi: 10.31098/lomr.v3i2.2824
Pereira R. M., Casaroli D., Vellame L. M., Oliveira F. A., Rodrigues R. F., Pires A. L., et al. (2020). Water deficit detection in sugarcane using canopy temperature from satellite images. Aust. J. Crop Sci 14, 400–407. doi: 10.21475/ajcs.20.14.03.p1647
Picoli M. C. A., MaChado P. G., Duft D. G., Gomes L. M., Fernandes F. H., Leite J. E., et al. (2019). Sugarcane drought detection through spectral indices derived modeling by remote-sensing techniques. Model. Earth Syst. Environ. 5, 1679–1688. doi: 10.1007/s40808-019-00619-6
Rahman M. R., Islam A. H. M. H., Rahman. M. A. (2004). NDVI derived sugarcane area identification and crop condition assessment. Plan Plus 1, 1–12.
Ranjan R., Chopra U. K., Sahoo R. N., Singh A. K., Pradahan. S. (2012). Assessment of plant nitrogen stress in wheat (Triticum aestivum L.) through hyperspectral indices. Int. J. Remote Sens. 33, 6342–6360. doi: 10.1080/01431161.2012.687473
Reyes-Trujillo A., Daza-Torres M. C., Galindez-Jamioy C. A., Rosero-García E. E., Muñoz-Arboleda F., Solarte-Rodriguez. E. (2021). Estimating canopy nitrogen concentration of sugarcane crop using in situ spectroscopy. Heliyon 7, e06566. doi: 10.1016/j.heliyon.2021.e06566
Simoes W. L., Oliveira A. R., Tardin F. D., Brito J. S., de Souza A. P., Ribeiro. R. V. (2023). Saline stress affects the growth of Saccharum complex genotypes. J. Agron. Crop Sci 209, 613–618. doi: 10.1111/jac.12647
Singels A., Jarmain C., Bastidas-Obando E., Annandale. J. (2018). Monitoring water use efficiency of irrigated sugarcane production in Mpumalanga, South Africa, using SEBAL. Water SA 44, 636–646. doi: 10.4314/wsa.v44i4.12
Singels A., Van Den Berg M., Smit M. A., Jones S., Bezuidenhout P., Van Antwerpen. M. (2010). Modelling water uptake, growth, and sucrose accumulation of sugarcane subjected to water stress. Field Crops Res. 117, 59–69. doi: 10.1016/j.fcr.2010.02.003
Soltanikazemi M., Minaei S., Shafizadeh-Moghadam H., Mahdavian. A. (2022). Field-scale estimation of sugarcane leaf nitrogen content using vegetation indices and spectral bands of Sentinel-2: Application of random forest and support vector regression. Comput. Electron. Agric. 200, 107130. doi: 10.1016/j.compag.2022.107130
Som-Ard J., Atzberger C., Izquierdo-Verdiguier E., Vuolo F., Immitzer M. (2021). Remote sensing applications in sugarcane cultivation: A review. Remote Sens. 13, 4040. doi: 10.3390/rs13204040
Sundara B., Natarajan V., Hari. K. (2002). Influence of phosphorus solubilizing bacteria on the changes in soil available phosphorus and sugarcane and sugar yields. Field Crops Res. 77, 43–49. doi: 10.1016/S0378-4290(02)00048-5
Suresha G. S., Hari K., Murali P., Rajamani L., Kalaiselvi K., Gopalareddy K., et al. (2024). “Sugarcane Fibre as Dietary Supplement, Health Benefits and Applications in Functional Foods,” in Value Addition and Product Diversification in Sugarcane. Eds. Suresha G. S., Krishnappa G., Palanichamy M., Mahadeva Swamy H. K., Kuppusamy H., Govindakurup H. (Springer, Singapore). doi: 10.1007/978-981-97-7228-5_17
Swami P., Anusha M., David K. M., Saivamsireddy G., Upadhyay L., Panigrahi C. K. (2025). Innovations in remote sensing techniques for monitoring agricultural crops and environmental stress conditions: A review. J. Sci. Res. Rep. 31, 178–194. doi: 10.9734/jsrr/2025/v31i63120
Tayade A., Vasantha S., Anusha S., Kumar R., Hemaprabha. G. (2020). Irrigation water use efficiency and water productivity of commercial sugarcane hybrids under water-limited conditions. Trans. ASABE 63, 125–132. doi: 10.13031/trans.13548
Teixeira A. D. C., Leivas J. F., Ronquim C. C., Ferreira L. G., Moreira E. F., dos Santos. J. P. (2016). Sugarcane water productivity assessments in the São Paulo State, Brazil. Int. J. Remote Sens. Appl. 6, 84–95.
Triadi R., Herdiyeni Y., Tarigan. S. D. (2020). “Estimating crop water stress of sugarcane in Indonesia using Landsat 8,” in 2020 International Conference on Computer Science and Its Application in Agriculture (ICOSICA). 1–4. Bogor, Indonesia. doi: 10.1109/ICOSICA49951.2020.9243255
Vasantha S., Gomathi R. (2012). Growth and development of sugarcane under salinity. J. Sugarcane Res. 2, 1–10.
Veysi S., Naseri A. A., Hamzeh S., Eftekhari. A. (2017). A satellite-based crop water stress index for irrigation scheduling in sugarcane fields. Agric. Water Manage. 189, 70–86. doi: 10.1016/j.agwat.2017.04.016
Veysi S., Naseri A. A., Hamzeh S., Mohammadzadeh A., Karami M., Mojaradi. B. (2020). Relationship between field measurement of soil moisture in the effective depth of sugarcane root zone and extracted indices from spectral reflectance of optical/thermal bands of multispectral satellite images. J. Indian Soc. Remote Sens. 48, 1035–1044. doi: 10.1007/s12524-020-01135-2
Vinayaka, Prasad P. R. C. (2024). AI-enhanced remote sensing applications in Indian sugarcane research: A comprehensive review. Sugar Tech 26, 609–628. doi: 10.1007/s12355-024-01409-w
Virnodkar S. S., Pachghare V. K., Patil V. C., Jha. S. K. (2020). Remote sensing and machine learning for crop water stress determination in various crops: A critical review. Precis. Agric. 21, 1121–1155. doi: 10.1007/s11119-020-09711-9
Virnodkar S. S., Pachghare V. K., Patil V. C., Jha. S. K. (2021). Denseresunet: An architecture to assess water-stressed sugarcane crops from Sentinel-2 satellite imagery. Trait du Signal 38, 1131–1139. doi: 10.18280/ts.380424
Vu N.-T., Bui T.-K., Vu T.-T.-H., Nguyen T.-H., Le T.-T.-C., Tran A.-T., et al. (2023). Biochar improved sugarcane growth and physiology under salinity stress. Appl. Sci. 13, 7708. doi: 10.3390/app13137708
Waqas M. M., Awan U. K., Cheema M. J. M., Ahmad I., Ahmad M., Ali S., et al. (2019). Estimation of canal water deficit using satellite remote sensing and GIS: A case study in Lower Chenab Canal System. J. Indian Soc. Remote Sens. 47, 1153–1162. doi: 10.1007/s12524-019-00977-9
Watanabe K., Agarie H., Aparatana K., Mitsuoka M., Taira E., Ueno M., et al. (2022). Fundamental study on water stress detection in sugarcane using thermal image combined with photosynthesis measurement under a greenhouse condition. Sugar Tech 24, 1382–1390. doi: 10.1007/s12355-021-01087-y
Waters E. K., Chen C. C. M., Azghadi M. R. (2025). Sugarcane health monitoring with satellite spectroscopy and machine learning: A review. Comput. Electron. Agric. 229, 109686. doi: 10.1016/j.compag.2024.109686
Wiedenfeld R. P. (1995). Effects of irrigation and N fertilizer application on sugarcane yield and quality. Field Crops Res. 43, 101–108. doi: 10.1016/0378-4290(95)00043-P
Wood A. W., Muchow R. C., Robertson. M. J. (1996). Growth of sugarcane under high input conditions in tropical Australia. III. Accumulation, partitioning, and use of nitrogen. Field Crops Res. 48, 223–233. doi: 10.1016/S0378-4290(96)00043-3
Yang Y., Gao S., Jiang Y., Lin Z., Luo J., Li M., et al. (2019). The physiological and agronomic responses to nitrogen dosage in different sugarcane varieties. Front. Plant Sci 10. doi: 10.3389/fpls.2019.00406
Yang H., Shukla M. K., Mao X., Kang S., Du. T. (2019). Interactive regimes of reduced irrigation and salt stress depressed tomato water use efficiency at leaf and plant scales by affecting leaf physiology and stem sap flow. Front. Plant Sci 10. doi: 10.3389/fpls.2019.00160
Yueanket A., Phunklang S., Kornsing S., Kupimai M., Innok A., Pumpoung T., et al. (2024). “AIoT-based water management solutions for sustainable sugarcane farming,” in 2024 International Conference on Power, Energy and Innovations (ICPEI). 19–23. Nakhon Ratchasima, Thailand. doi: 10.1109/ICPEI61831.2024.10748644
Keywords: AI-driven remote sensing, climate change, crop water stress index, leaf nitrogen content, salinity stress, sugarcane agriculture
Citation: Vinayaka, Prasad PRC, Avinash G, Amaresh, Arun Kumar R, Murali P, Palaniswami C and Govindaraj P (2025) Harnessing AI and Remote sensing for precision sugarcane farming: tackling water stress, salinity, and nitrogen challenges. Front. Agron. 7:1681294. doi: 10.3389/fagro.2025.1681294
Received: 07 August 2025; Accepted: 21 October 2025;
Published: 05 November 2025.
Edited by:
Santosha Rathod, Indian Institute of Rice Research (ICAR), IndiaReviewed by:
Santosh Ganapati Patil, Tamil Nadu Agricultural University, IndiaRahul Patil, University of Agricultural Sciences Raichur, India
Copyright © 2025 Vinayaka, Prasad, Avinash, Amaresh, Arun Kumar, Murali, Palaniswami and Govindaraj. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Vinayaka, dmluYXlha2EuYjN2c0BnbWFpbC5jb20=
†These authors have contributed equally to this work