UTAR Institutional Repository

Plant-disease detection by using computer vision approach

Chang, Man Kien (2023) Plant-disease detection by using computer vision approach. Final Year Project, UTAR.

Download (7Mb) | Preview


    Plant maladies have long been a major concern in agriculture, frequently resulting in substantial yield losses, economic losses, and degraded crop quality. As the global demand for food security and sustainable agricultural practices increases, there is a pressing need for effective and precise disease detection mechanisms. Computer vision and deep learning provide promising avenues for the rapid and accurate identification of plant diseases. This study explores the feasibility of utilising pre-trained deep learning models, such as ResNet18, VGG16, AlexNet, and GoogleNet, to detect and classify a wide variety of plant diseases. Using a comprehensive dataset containing images of foliage exhibiting various disease symptoms, these models were trained, refined, and evaluated with extreme care. According to preliminary findings, GoogleNet outperforms its competitors in terms of accuracy and computational efficiency. While apple leaves serve as the study's primary case study, the methodologies and findings have broader implications. It paves the way for the development of real-time disease detection systems on the field, which could revolutionise the agricultural industry. Such systems could endow farmers around the world with the means to make informed decisions, optimize crop health, and ultimately increase food production.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: S Agriculture > SB Plant culture
    T Technology > TK Electrical engineering. Electronics Nuclear engineering
    Divisions: Faculty of Engineering And Green Technology > Bachelor of Technology (Honours) in Electronic Systems
    Depositing User: ML Main Library
    Date Deposited: 01 Jan 2024 20:08
    Last Modified: 01 Jan 2024 20:08
    URI: http://eprints.utar.edu.my/id/eprint/6074

    Actions (login required)

    View Item