Cassava Plant Disease Detection on Low-End Mobile Devices Using Convolutional Neural Networks
Cassava is a staple crop used and consumed in predominantly Africa and other parts of the world. However, farmers lose significant portions of potential crop yield due to various cassava plant diseases. Although farmers use fertilisers and other preventive measures to mitigate these diseases, they are usually expensive and hence not very effective and predictive of future diseases. Convolutional Neural Networks(CNNs) provide an avenue to tackle this issue by providing a model that helps with predictive diagnosis of these diseases with near precise accuracy, even for low-income farmers. This project employed CNNs to build a classification model for cassava plant diseases with an accuracy of 89%. The model was then transformed to work with low-processor power devices and fed to the backend of an Android application. It demonstrated the efficacy of using Machine Learning techniques to develop applications for low-income farmers that can impact crop yield and agriculture in general.
Capstone Project submitted to the Department of Engineering, Ashesi University in partial fulfillment of the requirements for the award of Bachelor of Science degree in Computer Engineering, May 2022
Philip Nii-Otu Narteh. (2022). Cassava Plant Disease Detection on Low-End Mobile Devices Using Convolutional Neural Networks. Ashesi University.