An Innovative Hybrid Feature Extraction Method for the Diagnosis of Coconut Leaf Diseases

In South India, cultivating coconuts is a major
agricultural activity, but the trees are challenged by
unfavorable weather conditions and environmental factors.
These difficulties comprise pest infestations and a variety of
leaf diseases. Because of the coconut trees' enormous foliage
and shadowing, it might be challenging to identify and locate
these problems. The study proposes a feature extraction
strategy which integrates the feature vectors of deep learning
(DL) and handcrafted features into a single, cohesive
representation for detecting the coconut leaf diseases. The
selected features were fed into the Machine learning (ML)
classifier like Support Vector Machine (SVM), Random Forest
(RF) and XGBoost. The experimental results show that the
proposed technique outperforms with an accuracy of 97.4 %
showing the importance of feature extraction techniques in
identifying the coconut leaf diseases.
Keywords— : Coconut leaf disease detection, Feature
Extraction. Handcrafted features, Deep Features, Classifiers