| dc.description.abstract | Cinnamon is one of the most essential spice
crops with significant global economic importance,
and Sri Lanka is renowned for producing the highest
quality cinnamon, known as Ceylon Cinnamon.
Contributing to over 90% of the global spice market,
cinnamon is the most exported spice from Sri Lanka,
particularly to countries such as Mexico, the U.S.A,
and Peru. Between 2017 and 2022, cinnamon exports
increased by 1,618 metric tons, and its export value
grew by Rs. 40,843 million. However, diseases such as
leaf spot, rough bark disease, and stripe canker
present significant threats to cinnamon cultivation,
impacting both yield quality and quantity. This
research aims to address these challenges by
developing a machine learning and image processingbased system for the early detection of these three
prevalent cinnamon diseases. The system not only
identifies the diseases but also provides detailed
information on symptoms and treatments, improving
disease management for farmers. A dataset of
diseased cinnamon plant images, provided by the
National Cinnamon Research Center in Matara, was
used for training, consisting of leaf spot, rough bark,
stripe canker and healthy image samples. The
DenseNet-121 model was employed to train the
system, which achieved a high accuracy rate of 94%.
This system has the potential to significantly mitigate
the adverse effects of these diseases, enhancing both
productivity and the quality of cinnamon quills.
Additionally, it offers guidance to cinnamon peelers by
providing a peeling guide to support the peeling
process. Ultimately, this study aims to enhance
cinnamon production through timely disease detection
and effective treatment strategies. | en_US |
