Agripedia: a deep learning approach for crop health monitoring using yolov5 and vgg-16-based leaf disease detection/
Jhan Kyle V. Agullo, Deniece Winslhet A. Gabaon, John Patrick I. Marasigan, and Matthew A. Navale.--
- Manila: Technological University of the Philippines, 2025.
- vii, 127pages: 29cm:
Bachelor's thesis
College of Science.--
Includes bibliographic references and index.
Agriculture is a vital sector in the Philippines, employing 25% of the workforce and contributing 8.9% to the GDP. However, challenges such as climate change, pests, and limited access to expert knowledge hinder productivity, especially among small-scale farmers. To address this, Agripedia–an IoT-enabled crop monitoring system–was developed to promote backyard farming by providing real-time environmental tracking and AI-driven disease detection. The system integrates Arduino-based sensors (temperature, humidity, soil moisture and light intensity) with the help of a camera to assess crop health, an object detection model: YOLOv5 which detects leaves from images captured on the camera. while a convolutional neural network (CNN) analyzes leaf images to identify diseases. The mobile app processes this data, offering actionable recommendations to optimize yield and prevent crop loss. Agripedia also features a crop database with planting guides, tailored for Filipino users to enhance food security during crises like pandemics and typhoons. The system was evaluated through demonstrations and surveys distributed to 30 respondents (backyard farmers, agricultural workers, and IT professionals) based on ISO 25010 standards. Results showed strong agreement in functionality, usability, reliability, and maintainability. By democratizing access to precision farming tools Agripedia aims to empower communities, improve crop resilience, and contribute to sustainable agriculture in the Philippines.