AI-Powered Digital Monitoring and Decision Support Model for Smart Farming Applications

Abstract

The increase in global population, the impact of climate change on agricultural production, and the limited availability of agricultural land necessitate increasing agricultural productivity and developing digital monitoring systems focused on smart farming. Maintaining genetic purity in seed production is particularly important for crop quality and economic sustainability. However, traditional manual field inspections are insufficient in terms of time, labor, and accuracy across large areas. This study develops a digital monitoring and decision support model integrating AI-powered drones and mobile systems to monitor agricultural production areas and examines its effectiveness in field conditions. The application scenario focuses on monitoring isolation boundaries in seed sunflower production areas and detecting tassels to prevent foreign pollen contamination in seed maize production. The high-resolution aerial images obtained are analyzed using deep learning-based object recognition and pattern recognition algorithms YOLOv5, integrated into mobile and web-based decision support platforms. Field applications demonstrate that the developed model achieves accuracy between 85-96% under different crops and scenarios, maintains low false negative rates, and completes inspections that take hours or days with manual methods in minutes. In addition, approximately 40% of cost savings were achieved in operational processes. The results show that AI-powered digital monitoring systems offer an effective smart farming solution in terms of technical reliability, operational efficiency, and sustainability in agricultural production.