This thesis addresses critical challenges in Thai agriculture, particularly within greenhouse farming, by developing an innovative, cost-effective, and user-friendly automation solution for greenhouse management, leveraging Artificial Intelligence (AI) and Internet of Things (IoT) as well as drone technology to enhance agricultural productivity in Thailand.

Inspired by insights from agricultural experts and challenges identified during expert interviews—including technological limitations, inefficient resource allocation, and the demand for automation—this research aims to empower farmers with accessible tools and technologies to enhance resource allocation, improve crop quality, and promote sustainable agricultural practices.

The proposed system integrates AI for predictive decision-making and autonomously adjusts watering schedules and fertilizer dosages based on real-time environmental data and crop growth stages, utilizing advanced Computer Vision algorithms to analyze drone-captured imagery. By determining crop stages from seedling to fruit-bearing, the system optimizes resource use and streamlines labor planning.

Through extensive evaluations and fine-tuning of YOLOv8 models for real-time tomato recognition in greenhouse environments, this study highlights the model’s speed and adaptability, emphasizing the critical role of learning rate optimization and dataset quality in achieving optimal accuracy.

The findings demonstrate that the optimized YOLOv8n model can significantly enhance agricultural monitoring efficiency, contributing to the ongoing dialogue on innovative solutions in the agricultural sector and fostering improved crop quality and sustainable practices. Ultimately, this research provides practical solutions for sustainable farming, addressing pressing needs in the broader field of Precision Agriculture and setting the stage for future advancements.