Disease detection of tomato crop in a hydroponic farming system using a convolutional neural network

Journal Title
Journal ISSN
Volume Title
There is a need to explore hydroponic farming as a new way of farming produce in urban areas in Ghana. Hydroponic farms usually take up less space than traditional farms and employ a level of automation, making them easier to manage and a more appealing option for the youth. Crop disease identification is usually made best with expert eyes; however, work has been done to improve disease detection using computer vision on single leaves. To aid in the automation of disease identification, there is the need to detect and classify crop diseases on leaf clusters to implement preventions early and increase crop yield. This project focuses on detecting five tomato crop diseases, namely, Bacterial spot, Black leaf mould, Grey leaf spot, Late Blight and Powdery mildew diseases. The resulting model has an accuracy of 99.01% on the test dataset, which included images of single tomato leaves, clustered tomato leaves and images with varying brightness. The model is then uploaded onto Android and iOS mobile devices for convenient, real-time disease classification. The accuracy obtained from the mobile devices is about 90% ± 6.8% when using images from the internet.
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
automated agriculture