This work presents a conceptual design and system architecture for a mobile-based agricultural diagnostic platform aimed at bridging the technological gap between large-scale commercial farming and smallholder farmers. The proposed system leverages computer vision and on-device machine learning to deliver real-time crop disease detection, nutrient deficiency analysis, and treatment recommendations directly through a smartphone camera — without requiring specialized agricultural hardware. The system is built around a convolutional neural network (CNN) trained on the PlantVillage dataset, deployed via TensorFlow Lite for lightweight on-device inference. Key design priorities include offline functionality, multilingual support, and low-bandwidth operation, making the tool viable in rural, low-connectivity environments. A hybrid decision-support module combines ML outputs with an agronomic rule database to generate localized treatment guidance in both organic and conventional farming contexts. The paper includes a review of related literature on CNN-based plant pathology detection, a comparative analysis of existing applications (Plantix, PlantVillage Nuru, CropIn, OneSoil), system requirements, and UML diagrams covering the sequence, block, and use-case views of the proposed architecture.This work was conducted at Arab International University (AIU), Syria. The official website of the university is: https://www.aiu.edu.sy
swadi et al. (Mon,) studied this question.