Automated Quantitative Assessment of Apple Scab Disease Severity Using Low-Cost IoT Camera Networks and Machine Learning
Faculty Mentor Information
Dr. Mary Everett, University of Idaho
Presentation Date
7-16-2026
Abstract
Apple scab is a fungal disease that reduces orchard yield and fruit quality, and effective management requires consistent, long-term monitoring of its progression across the canopy. This project is the foundational phase of a larger effort to develop a low-cost IoT camera network paired with machine learning for automated, quantitative assessment of apple scab severity. Given the scope of this grant period, work focused on designing and building the data collection hardware that will ultimately feed this system. The device integrates a custom 3D-designed weatherproof enclosure, a solar power system with regulated power conversion, a Raspberry Pi Zero 2 W single-board computer for image capture and data transmission, and an ESP32 microcontroller that manages power scheduling by safely powering the Pi on and off to conserve energy between capture cycles. The result is a functional, deployable device that autonomously photographs tree canopies and logs temperature and humidity data on a set schedule. This device serves as a scalable data collection point; deploying multiple units across an orchard will generate the image and environmental datasets needed for the machine learning-based severity analysis planned in future work.
Automated Quantitative Assessment of Apple Scab Disease Severity Using Low-Cost IoT Camera Networks and Machine Learning
Apple scab is a fungal disease that reduces orchard yield and fruit quality, and effective management requires consistent, long-term monitoring of its progression across the canopy. This project is the foundational phase of a larger effort to develop a low-cost IoT camera network paired with machine learning for automated, quantitative assessment of apple scab severity. Given the scope of this grant period, work focused on designing and building the data collection hardware that will ultimately feed this system. The device integrates a custom 3D-designed weatherproof enclosure, a solar power system with regulated power conversion, a Raspberry Pi Zero 2 W single-board computer for image capture and data transmission, and an ESP32 microcontroller that manages power scheduling by safely powering the Pi on and off to conserve energy between capture cycles. The result is a functional, deployable device that autonomously photographs tree canopies and logs temperature and humidity data on a set schedule. This device serves as a scalable data collection point; deploying multiple units across an orchard will generate the image and environmental datasets needed for the machine learning-based severity analysis planned in future work.