SUNFLOWER DISEASE PREDICTION USING ARTIFICIAL INTELLIGENCE IN PERCEPT OF BANGLADESH
No Thumbnail Available
Date
2022-11-30
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Plant disease prediction and recognition in the early stage is one of the most essential needs to
increase agriculture, which plays an important role in our country's economy and helps to feed
a large population. Sunflower is a plant categorized as a low to medium drought-sensitive crop.
But now a days sunflower production faced a sever crisis due to its many diseases. Farmers
also lost their interest to the production of sunflower. But if proper action is received earlier,
many serious disease will have no effect on the plants. It also improve the productively,
quantity and quality of sunflower plants. Manual identification of sunflower disease is an
exhausting task at time. Now Artificial intelligent is a great field and gained its popularity for
the prediction of plant disease as early as possible. In this research paper different machine
learning techniques such as Linear Regression (LR), Random Forest (RF), Gradient Boosting
(GB), Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Extreme Gradient
Boosting (XGBoosting) are applied in the image features that are extracted using digital image
processing technique. And deep learning techniques such as Residual Network 50 (ResNet 50),
Visual Geometry Group 19 (VGG 19) and Mobile Net Version 2 (MobileNet V2) are used in
the direct image for predicting the disease of sunflower. A total of 602 images which are
collected from Bangladesh Agricultural Institute, gazipur is used in this research work.
Machine learning techniques GB shows 93.38 % accuracy on the other hand XGBoost Shows
92.56% accuracy but highest sensitivity 0.92. Deep learning model ResNet 50 Shows highest
accuracy and sensitivity 95.09 % and 0.95. Among Machine learning and Deep learning models
ResNet 50 shows highest accuracy and sensitivity, therefore ResNet 50 is the best model for
predicting sunflower disease for this dataset.