Published newsletter
Article Details
A groundbreaking PhD study using Artificial Intelligence (AI) to detect and assess fruit fly damage in mangoes was presented at the Adaptive Environment Monitoring Network For East Africa (AdEMNEA) Project’s Annual General Meeting (AGM) on November 13-14, 2025.
Presenting his work, Yonas Safari, a PhD student of the project in Work Package 3, explained that his research focuses on applying machine learning techniques to estimate the severity of damage caused by fruit flies - one of the most destructive pests affecting mango production. He noted that fruit flies can cause losses of up to 100% in severe cases.
While many previous studies have focused on post-harvest damage detection, Safari’s work addresses a major gap by targeting pre-harvest assessment, enabling farmers to understand damage levels while fruit is still in the orchard. Using phone-based RGB images and thermal imaging, his study captures thousands of fruit images for analysis.
The collected data is processed and fed into deep learning models that automatically detect the severity of damage. The system distinguishes healthy areas from damaged ones and accurately estimates loss levels. One of the studies from this work, using machine learning models, has already been published online.
Safari further revealed that thermal imaging plays a critical role in detecting internal damage that cannot be seen with ordinary images.