To fulfill the rising need for food, water, and energy, increasing socioeconomic development and population increase, resulting in new problems related to agriculture. To effectively safeguard the sustainability of resources and manage the impact of future agricultural output, rethinking is needed. By employing a combined energy-water-food connection optimization technique, it has been found that it is currently not possible to manage agricultural resources and waste in accordance with agricultural management principles to achieve sustainability. ¹

The advancement of agricultural modernization is strongly supported by agricultural information. The implementation of modern technological advances to provide relevant and helpful data to users through agricultural information suggestion services has been shown to be a successful remedy with the ongoing evolution of agricultural information infrastructure construction. ² This promoted the creation of autonomous farming systems with a primary focus on irrigation, climate control, and greenhouse monitoring for field work, animal management, and growth control. Effective agricultural production management, which increases both the productivity and safety of agricultural products, can reduce negative environmental effects. Technologies based on sensors are essential in the expanding field of agriculture. Farmers, scientists, and technological manufacturers are working together to develop practical solutions that will increase productivity and result in cost savings. ³

Agriculture depends on numerous economic and climatic conditions. The variables that affect agricultural crop production include soil, weather, cultivation, temperature, irrigation, fertilizers, rainfall, harvesting, pesticide and insecticide usage, weeds, and other variables. The study of vital phenomena in plants is known as plant physiology. Understanding different plant biological processes, such as photosynthesis, respiration, transpiration, translocation, nutrient uptake, hormone-regulated plant growth, and other activities that have a significant impact on crop yield, is helpful. ⁴

The profound integration of both contemporary information technology and conventional farming has given rise to the “smart agriculture” era. Smart agriculture addresses automation and intelligence in agriculture. ⁵ Nonetheless, the concerns of statistical security cannot be ignored considering the growth in agriculture facilitated by modern digital technologies. In agriculture, statistical analysis has an important impact on the collection, analysis, and interpretation of numerical data. Agriculturalists can utilize data analytics to develop predictive analytics for future yields, manage resources based on existing trends, and continuously check the health of their crops, ⁶ maximizing profitability and minimizing waste. Statistical analysis of agricultural parameters helps to manage agricultural resource growth by identifying optimal conditions for various crops. Some crops may require a specific range of temperature, humidity, and soil moisture for optimal growth. Understanding the optimal conditions for crop growth, farmers can apply these resources in a targeted manner, reducing waste and maximizing their impact.

The utilization of IoT devices for monitoring and controlling crop production is prevalent, ⁷ but rural areas often face challenges due to unstable cloud connectivity. Fog computing technology addresses this issue by managing sporadic connectivity and enabling faster data analysis. The experiment employed two datasets: one with air humidity and temperature values and another with soil temperature and moisture values. Fog filters apply unsupervised machine learning techniques to cluster unlabeled data and use supervised learning classification techniques to predict data sample categories.

A comprehensive analysis of 61 selected studies from 221 publications focused on advanced time series models in agriculture. ⁸ Historical data and information techniques for applying time series models are discussed. This review highlighted that deep neural networks enhance simulation and data structure mining capabilities, contributing to end-to-end optimization for environmental prediction. This approach broadens the range of environmental characteristics that agricultural facilities can manage.

A systematic review of agricultural IoT presented the current state, system architecture, and five main IoT technologies. ⁹ Five chosen fields for IoT applications in agriculture were introduced, along with an analysis of the challenges and future growth predictions. The key issues identified include network and interoperability problems, trust, environmental sustainability, and insecurity. An evaluation of IoT technology in agri-food supply chains identified 24 crucial criteria categorized into technological, social, economic, and organizational aspects. ¹⁰ The DEMATEL approach was used to determine causal relationships, highlighting trust, environmental sustainability, insecurity, interoperability, and network issues as significant factors influencing IoT adoption. These insights can help overcome obstacles in the agri-food sector.

Research on tomato farming in Mayotte aimed to understand production practices and technical choices to promote environmentally friendly practices. ¹¹ Field inspections and farmer interviews revealed the overapplication of pesticides and a lack of correlation between application rates and crop health. This study indicated a need for better information and practices related to agricultural health protection.

Data mining techniques such as clustering and regression were used to analyze agricultural data, identifying optimal parameters for enhancing crop production. ¹² This approach helps improve productivity and climate change resistance by analyzing new and existing data on crops, soil, and climate. An agricultural management system was designed to increase productivity and profitability through effective practices and online product sales. ¹³ Data fusion techniques combine information from various sources for accurate remote findings, aiding intelligent decision-making for resource management based on seasonal and environmental conditions.

The state of modern farm management systems was assessed, highlighting each stage from data collection to decision-making. ¹⁴ Using AI and robotic technologies, data-driven managed farms can enhance productivity and reduce environmental harm, establishing a foundation for sustainable agriculture.

Agricultural development data from government publications were analyzed using correlation and multiple regression to understand the relationships between cropping intensity and various production performance factors. ¹⁵ The study revealed significant variance in crop yields influenced by factors such as irrigation and climate conditions. Big Data Analytics was used to construct a weather-based crop prediction system in India. ^{16, 17} Data on various climatic and soil factors were preprocessed and analyzed using the MapReduce framework and k-means clustering. The system provided accurate predictions and was presented through a graphic user interface, aiding agriculturalists in optimizing yields. Systems thinking was implemented to improve agricultural land production by using an internet-based causal loop diagram. ¹⁸ This approach helps measure variables such as temperature, humidity, and pest activity, addressing issues in planting and pest control.

A comprehensive examination of data mining software in agriculture classified methods for crop production monitoring. ¹⁹ This study suggested an intelligent crop management system to help farmers make informed decisions based on various environmental and crop parameters. Understanding soil water and salinity variations is crucial for effective irrigation and fertilization management. ² Studies have provided 48-hour estimates of soil salt and water levels, aiding prediction and management in semiarid regions. ²⁰ Effective planning and yield assessment are essential in agriculture. The use of data mining techniques for making critical farming decisions, considering commodity prices, soil variance, and environmental factors, was emphasized. Machine learning techniques such as random forest regressor and linear regression were used to analyze agricultural data and optimize crop productivity. ^{21, 22} To address the challenges of climate change and food production, climate-smart farming methods were compared with traditional methods. Irrigation and smart crop management techniques significantly increase yields, demonstrating the potential of data fusion models in optimizing agricultural practices. ^{23, 24}

This literature survey highlights the diverse applications of smart agriculture technologies, focusing on IoT, fog computing, data mining, machine learning, and climate-smart farming methods to enhance productivity, sustainability, and resource management in agriculture.

Despite advancements in statistical analysis tools for deciphering crop parameter levels, there remains a significant gap in understanding the inherent variability in crop phenology. Different crops exhibit unique growth stages and respond variably to environmental conditions, posing a challenge in developing generalized models. Research needs to focus on creating more crop-specific models that can accurately capture these unique phenological responses and environmental interactions.

The primary focus of analyzing agricultural parameters is twofold: to assist researchers in expanding their knowledge base and to ensure that farmers benefit from improved crop outcomes and effective field management. Through careful examination of the data, researchers can identify patterns, correlations, and trends, enabling them to make targeted recommendations for optimizing crop production. The vision of designing an autonomous crop management system that is both technologically advanced and flexible remains largely unfulfilled. Existing systems often lack the necessary adaptability to cater to the diverse requirements of different crops and their unique phenologies. Research should focus on developing more flexible, adaptive systems that can tailor interventions and recommendations specific to each crop and its environmental conditions. In conclusion, ongoing advancements in agricultural research, coupled with the application of sophisticated analysis tools, are instrumental in ushering in a new era of precision farming. The goal is to design autonomous systems that not only streamline farming operations but also adapt intelligently to the nuanced needs of different crops, thereby fostering a more sustainable and productive agricultural landscape.

An agricultural crop management system refers to the combination of various technologies, such as sensors, robotics, and artificial intelligence, to automate and optimize the tasks involved in crop cultivation. Statistical analysis plays a crucial role in agricultural crop management systems. In this work, the capsicum crop growth ²⁵ experimental dataset was used for the analysis. The dataset consists of agricultural parameters such as temperature, humidity, and soil moisture (at three different levels, SM1, SM2, and SM3) and pressure and light intensity recorded for three months. The Capsicum crop dataset available was recorded during the months of March 2020 (328 samples), April 2020 (2880 samples), May 2020 (2976 samples) and June 2020 (1004 samples) using Libelium hardware. In this work, the temperature, humidity, and soil moisture (SM1) datasets are considered for statistical analysis.

The dataset was analyzed and compared with the available phonological (growth) stage data of Capsicum crops. ^{26, 27} Surveying and consulting agriculturalists revealed that the phenological stages of Capsicum consisted of four growth stages, as shown in Table 1. The data were further divided into four phenological stages.

The capsicum seeds are initially seeded in trays or cavities. Seeds germinate approximately one week after sowing. The seedlings in trays are transplanted to fields or planting beds within 30-35 days. ¹⁴ After transplanting the plants to the vegetative stage followed by flowering, the fruit set and harvest stages of the Capsicum plants were evaluated. The experimental dataset was large, and temperature, humidity, and soil moisture data were available for different growth stages of Capsicum. Box charts are useful visual tools for summarizing and analyzing large datasets. Figure 1 shows that the box plot of temperature data for crop growth stages provides insights into central tendencies, variability, and outliers, which helps to guide decisions related to crop management and identify potential areas for improvement.

A box plot of the humidity data is shown in Figure 2, which provides a visual representation of the humidity variability and helps to identify typical, extreme, and unusual conditions. A box plot of the soil moisture data is shown in Figure 3. It helps to assess and understand the soil moisture patterns and identify areas that might require attention or action. The methodology for conducting the statistical analysis of the agricultural parameters of temperature, humidity and soil moisture is shown in Figure 4. Statistical analysis is crucial for understanding and optimizing crop growth and farm management. In this work, the temperature, humidity, and soil moisture dataset of Capsicum was used for statistical analysis. Through surveying and consulting agriculturalists, the Capsicum dataset was divided into four phenological stages. Understanding plant phenology is important for designing a crop management system for adapting to changes in climate and ecological interactions. The min–max normalization technique is used to normalize the data as part of the preprocessing stage.

Statistical hypothesis tests were performed to determine whether there were significant differences or effects related to the parameters. A significance level of α = 0.05 was used to compare the computed p values. A p value of less than 0.05 indicates statistical importance and rejection of the null hypothesis. Statistical analysis is vital in autonomous crop management systems because it enables data-driven decision making and optimization of agricultural resources. Agriculture is increasingly utilizing data for efficient decision-making, enhancing productivity and sustainability. Smart farming is being driven by advancements in data management, utilizing objective information acquired through sensors to minimize resource misuse and environmental pollution. Data-driven agriculture, which involves the incorporation of robotic solutions and artificial intelligence techniques, aims to transform food production to meet population growth needs while saving money.

This study’s main goal was to assess how agricultural factors influence the growth and development of Capsicum plants throughout their phenological stages. To investigate this, the t test was employed, a statistical tool that allows for the comparison of means between two or more groups. The primary objective was to determine whether there are statistically significant differences in these agricultural parameters across various phenological stages of Capsicum crop cultivation. The mean and standard deviation (SD) values of the temperatures during the different phenological stages are shown in Table 3. These statistics provide a quantitative understanding of temperature variations throughout crop growth stages for t tests. The mean and standard deviation (SD) values of humidity during the different phenological stages are displayed in Table 4. This data table provides insights into the humidity levels at each stage.

The mean and standard deviation (SD) values of soil moisture during the different phenological stages are displayed in Table 5. This information can be valuable for making decisions related to crop management and optimizing irrigation practices. The t test analysis of the agricultural parameters of temperature, humidity, and soil moisture across different phenological stages of Capsicum crops provided compelling evidence to support the rejection of the null hypothesis. The t test results are tabulated in Table 6, and Figure 5 shows the t test results.

With a p value less than the chosen level of significance of 0.05, there were significant differences in the agricultural parameters between the specified phenological stages of the Capsicum crop. These findings underscore the importance of monitoring and managing these parameters throughout the growth cycle of Capsicum plants to optimize their growth and yield. Farmers and researchers can use this valuable insight to make informed decisions and implement targeted strategies for Capsicum cultivation, ultimately contributing to improved agricultural practices and crop productivity.

Analysis of variance (ANOVA) was used to evaluate the variations in temperature, humidity, and soil moisture across the various developmental stages of the Capsicum crop. A significance level (α) of 0.05 was selected as the threshold for determining statistical significance. The ANOVA results are organized and presented in Table 4 for clear reference and interpretation. In this study, we investigated whether there are significant differences in temperature, humidity, and soil moisture levels across different stages of Capsicum crop growth. Table 7 serves as a crucial resource for summarizing the outcomes of the ANOVA. It likely includes key statistical measures such as F values and p values corresponding to each of the three agricultural parameters (temperature, humidity, and soil moisture). These values are essential for making informed decisions regarding the acceptance or rejection of the null hypothesis.

The results of ANOVA performed on agricultural parameters for different stages of the Capsicum crop are shown in Figure 6. The obtained p values for soil moisture, temperature and humidity were 2.22045e-16, 0 and 1.22125e-15, respectively. The p value is less than the level of significance (α=0.05). There were significant differences among the groups in terms of these parameters. The use of ANOVA in this context enables agricultural scientists, researchers, and farmers to gain valuable insights into how temperature, humidity, and soil moisture influence the various stages of Capsicum crop development. Such knowledge can inform cultivation strategies, help to optimize environmental conditions and ultimately enhance crop yield and quality.

The formidable challenge faced by farmers and agriculturalists in analyzing vast amounts of agricultural crop data for resource utilization forecasting has been addressed through the application of open-source statistical tools. The anticipated productivity parameters were identified, contributing to efficient agricultural resource management. Our study specifically scrutinized variations in temperature, humidity, and soil moisture, focusing on Capsicum crop parameters. Through the application of statistical tools such as t tests and ANOVA, significant differences in the parameters were identified across the various phenological stages. The results indicated that temperature, humidity, and soil moisture levels varied notably between stages, underscoring the importance of precise monitoring and management of these factors to optimize crop growth and yield. The use of box plots facilitated the visualization of data variability and outliers, further aiding in decision-making processes.

The statistical findings are instrumental for the development of autonomous crop management systems that can make data-driven decisions. These systems are vital for improving agricultural resource management, enhancing productivity, and promoting sustainability. By leveraging statistical analysis, farmers and researchers can better understand the impacts of climate change and other variables on crop phenology, enabling the adaptation of cultivation practices to changing environmental conditions.

In summary, our research endeavors strive to offer a comprehensive solution to the intricate challenges faced by farmers, providing a pathway to improve crop yield and effective agricultural resource management through the development of an autonomous crop management system.

Conceptualization – Santhosh KV and Deepashri KM; methodology - Deepashri KM, Santhosh KV and J Satheesh Kumar; software – Deepashri KM and Santhosh KV; validation – Santhosh KV and J Satheesh Kumar; formal analysis – Deepashri KM; investigation, Deepashri KM and Santhosh KV; writing original draft preparation – Deepashri KM and J Satheesh Kumar; writing review and editing – Santhosh KV and J Satheesh Kumar; supervision – Santhosh KV.

Ethics and consent were not required.