THE DESIGN AND EVALUATION OF AN AI-BASED CROP DISEASE DETECTION SYSTEM: A CASE STUDY OF MAIZE FARMERS IN KADUNA STATE

THE DESIGN AND EVALUATION OF AN AI-BASED CROP DISEASE DETECTION SYSTEM: A CASE STUDY OF MAIZE FARMERS IN KADUNA STATE

CHAPTER ONE: INTRODUCTION

1.1 Background of the Study

Agriculture remains a cornerstone of Nigeria’s economy, contributing approximately 24% to the nation's GDP and employing over 70% of the rural labor force (National Bureau of Statistics [NBS], 2023). Among staple crops, maize is of critical importance due to its role in food security, animal feed, and as a raw material in agro-industrial processes. However, one of the persistent challenges undermining maize production is the prevalence of crop diseases, which significantly reduce yield and increase the vulnerability of smallholder farmers. In Kaduna State, where maize farming is widely practiced, disease outbreaks such as Maize Lethal Necrosis and Northern Leaf Blight continue to threaten sustainable agricultural productivity (Adebayo & Okafor, 2021).

Early and accurate detection of crop diseases is essential for timely intervention and disease management. Traditionally, disease detection relies heavily on manual inspection by farmers or agricultural extension officers. This method, however, is often limited by a lack of expertise, delayed responses, and misdiagnosis. Furthermore, extension services in many rural communities are overstretched and underfunded, making it difficult to reach a significant number of farmers in need of technical support (Oladele, 2022). These gaps have created an urgent need for innovative, scalable, and automated solutions that can empower farmers to identify crop diseases accurately and promptly.

Recent advances in Artificial Intelligence (AI), particularly in computer vision and machine learning, offer promising avenues for addressing this challenge. AI-powered systems can be trained to recognize visual symptoms of plant diseases through image analysis, enabling real-time detection using mobile devices or embedded sensors (Patel et al., 2020). These systems have the potential to democratize access to agronomic knowledge, reduce diagnostic errors, and enhance precision agriculture practices. In low-resource settings such as Kaduna State, the integration of AI-based solutions could significantly improve disease management outcomes and farmer productivity.

Despite global advancements, the adoption of AI in Nigerian agriculture is still in its nascent stages. There is limited empirical evidence on the effectiveness, usability, and acceptance of such technologies among local farmers. Furthermore, existing AI systems are often trained on foreign datasets, which may not accurately represent the visual characteristics of diseases on local maize varieties. This creates a critical gap in localized AI model development and evaluation. Therefore, there is a pressing need to design, develop, and assess a context-specific AI-based crop disease detection system tailored to the needs of maize farmers in Kaduna State.

1.2 Statement of the Problem

Crop disease remains a major impediment to maize production in Kaduna State, often resulting in substantial yield losses and economic hardship for smallholder farmers. Traditional methods of disease detection—based on visual observation and reliance on limited agricultural extension services—are insufficient in addressing this growing problem. Misdiagnosis, delayed intervention, and lack of timely information often exacerbate disease outbreaks, undermining food security and farmer livelihoods (Ibrahim & Musa, 2022).

Although Artificial Intelligence holds significant promise for transforming agricultural diagnostics, its practical application in Nigeria faces several challenges. First, there is a scarcity of AI models trained on locally relevant datasets. Most AI tools are developed using images of plant diseases from different climatic and agricultural contexts, reducing their accuracy and reliability in the Nigerian setting. Second, the usability and accessibility of AI solutions by smallholder farmers—many of whom have limited digital literacy—remain uncertain. Third, there is inadequate empirical evaluation of AI-based systems in field conditions to determine their effectiveness and limitations.

Without addressing these issues, the deployment of AI solutions in Nigerian agriculture risks being ineffective or unsustainable. This research seeks to fill these gaps by designing and evaluating an AI-based crop disease detection system that is tailored to local conditions, technically robust, and farmer-friendly. By focusing on maize farmers in Kaduna State, the study aims to contribute both technological innovation and contextual insight into AI-driven agricultural support systems.

1.3 Objectives of the Study

To design an AI-based system for the detection of common maize diseases in Kaduna State.

To evaluate the performance and accuracy of the AI system using locally collected data.

To assess the usability and acceptance of the AI-based system among maize farmers in Kaduna State.

1.4 Research Questions

What are the design requirements and technical features of an effective AI-based crop disease detection system for maize in Kaduna State?

How accurate and reliable is the AI system in detecting maize diseases under field conditions?

How do maize farmers in Kaduna State perceive and interact with the AI-based detection system?

1.5 Research Hypotheses

H0₁: The AI-based crop disease detection system does not significantly improve the accuracy of maize disease diagnosis compared to traditional methods.

H0₂: There is no significant relationship between system usability and farmers' willingness to adopt the AI-based detection tool.

1.6 Significance of the Study

This study is significant for multiple stakeholders. For researchers and technologists, it contributes to the development of AI models that are adapted to local agricultural contexts. For farmers, the study presents a potentially transformative tool that can improve disease diagnosis and crop management. For policymakers and agricultural extension agencies, the findings offer insights into the practical integration of digital technologies in rural farming systems. More broadly, the study supports national efforts to modernize agriculture and enhance food security through innovative digital solutions.

1.7 Scope and Limitation of the Study

The study is limited to maize farmers in Kaduna State and focuses specifically on the detection of common maize diseases using AI. It involves system design, technical evaluation, and user feedback. The study does not address broader issues of disease treatment, infrastructure limitations, or economic policy. Limitations may include variations in image quality, farmer bias during feedback collection, and challenges in long-term system deployment.

 

1.8 Operational Definition of Terms

Artificial Intelligence (AI): The simulation of human intelligence in machines, particularly for tasks like image recognition and decision-making.

Crop Disease Detection: The process of identifying signs and symptoms of plant diseases, particularly through visual analysis.

Machine Learning: A subset of AI that involves training algorithms to recognize patterns and make predictions based on data.

Usability: The ease with which users can interact with a system to achieve specific goals effectively and efficiently.

 

1.9 Structure of the Study

This study is structured into five chapters. Chapter One introduces the research problem, objectives, and scope. Chapter Two provides a review of related literature on AI applications in agriculture and crop disease detection. Chapter Three outlines the research methodology, including system design, data collection, and evaluation techniques. Chapter Four presents the findings and analysis. Chapter Five concludes the study with a summary, recommendations, and suggestions for future research.

 

References

Adebayo, S. A., & Okafor, C. N. (2021). Maize production challenges in Northern Nigeria: A review. Journal of Agricultural Science and Technology, 13(2), 105–117.

Ibrahim, Y. M., & Musa, H. G. (2022). Smallholder farmers and plant disease management in Northern Nigeria. African Journal of Rural Studies, 9(1), 77–89.

National Bureau of Statistics. (2023). 2023 Agricultural Sector Performance Report. Abuja: NBS.

Oladele, O. I. (2022). Agricultural extension and digital innovation: Challenges and prospects. Nigerian Journal of Agricultural Extension, 26(3), 119–131.

Patel, D., Singh, R., & Thomas, A. (2020). Deep learning-based plant disease detection: A review. Computational Agriculture, 5(1), 34–47.