Abstract:
Bangladesh is country of seasons. Naturally the people of Bangladesh are blessed
with various native seasonal fruits to fulfill their needs and appetites. So, hundreds of
local fruits are cultivated through the yearlong by the local farmers. But in recent
times some obsolete fruits like apple, blueberry, cherry, grape, orange, peach,
raspberry, strawberry etc. are getting popularity amongst local consumers of
Bangladesh and that is the reason why local farmers are now contemplating towards
the commercial cultivation of these obsolete fruits. But there are too many obstacles
in that process of commercial cultivation including weather demand, quality of soil,
cultivation technology, diseases, insects attack etc. Most of the time diseases of these
fruits notice our eyes when fruits are attacked and we cannot help but let the fruits get
rotten and suffer the financial loss too. Even though we might not handle every
barrier, but we can use deep learning and image processing technology for the
detection of diseases of these fruits and help the farmers taking necessary steps on
time. To simplify the work, we have taken the images of these obsolete fruits leaves
and implemented image processing algorithm and deep learning methods on them.
After the completion of this research we have accomplished an accuracy of 92.56%.
This research is going to help the farmers to cultivate and promote these obsolete
fruits more in a broaden way by reducing the diseases. It is an eco-friendly system
which detects diseases without much effort only by clicking attacked images and
putting that in the system and the system will give an output of which disease the
plant is attacked by.