| dc.description.abstract | Plants are considered an essential
part of our ecosystem and Sri Lanka has a
long history of using plants as a source of
medicines in Ayurveda. In addition to some
herbaceous plants serving as a food source,
have medicinal values. In the Ayurveda
medicinal industry, it is very important to
identify the correct herbs that help in the
preparation of remedial medicines. The
identification of these suitable herbaceous
plants is often done by skilled specialists.
However the problem is since identification
is based on human cognition, it can lead to
misjudgment. So it a waste that humankind
couldn’t use the herbal power of remedial
medications. To address this question the
paper proposes a simple and effectual
methodology for identification of Ayurveda's
herbaria, using mobile devices in the android
platform by implementing image processing
techniques. The main characteristics
required to identify a medicinal herb are the
shape, color, and texture of the leaf. The color
and texture of the leaf cover vital parameters
that are unique to a particular plant.
Preprocessing, feature extraction, and
classification are the three major phases in
the suggested methodology. In order to train
neural networks, images of herbal plant
leaves were captured under the supervision
of an Ayurveda doctor. For all the images
backgrounds are removed and resized before
applying classification techniques. According
to the methodology, the leaf images are
trained and the result can be shown through
the mobile application. The study got 94% of
accuracy for the proposed methodology | en_US |
