Title: Ai-Based Mobile App: Cash Crops Disease Detection In Smallholder Agriculture
Authors: Happy Sakuya , Italange Mpiguzi , Gift Mbugi , Elias Laurent , Esromu Erasto , Debora Isaya , Gerald Gabriel , Eva Saimon and Abdallah Ali
Volume: 10
Issue: 3
Pages: 14-19
Publication Date: 2026/03/28
Abstract:
Cash crops diseases cause significant yield losses in staple crops such as maize, tomato, and potato, threatening food security and smallholder farmers' livelihoods, especially in sub-Saharan Africa. Traditional detection methods visual inspection and laboratory analysis are slow, subjective, error-prone, and often inaccessible in rural areas. This leads to delayed action, excessive pesticide use, and environmental harm. Advances in smartphones and machine learning now enable mobile-based app for rapid and accurate leaf disease diagnosis. Farmers capture images in the field, and lightweight models provide instant results (disease name, confidence, severity) either on-device or via cloud. This review examines mobile machine learning approaches for Cash crops disease detection, including key architectures, datasets, and deployment strategies. It highlights high laboratory accuracy but significant performance drops in real-field conditions due to variable lighting, complex backgrounds, overlapping elements, and domain shift. The proposed Mobile-Based Cash crops Disease Detection System is evaluated against existing work, showing good alignment with smallholder needs while identifying gaps in robustness, offline functionality, and farmer-oriented features. Recommendations focus on realfield data collection, domain adaptation, offline inference, severity estimation, and user testing to support sustainable agriculture.