Enhanced Ensemble Learning and Feature Selection for Plant Disease Identification

Recognising plant diseases is essential for ensuring agricultural productivity, ensuring food security, and maintaining economic stability. The early and accurate identification of plant diseases enables timely intervention, which in turn contributes to the reduction of crop losses and the improvement of yield quality. The challenge is addressed by this research, which develops a hybrid approach to identify plant species and diagnose the health status of those species. CLAHE, is the first step in the process, which involves image enhancement. Following the extraction of a combination of deep learning-based and handcrafted features, the research proceeds to select the top 10% of features using Random Forest-based Recursive Feature Elimination (RFE), Extra Trees, and Select K-Best with Mutual Information. This is followed by a weighted selection of the features. A weighted majority-voting ensemble classifier is used to perform the classification and the performance of the model is validated through the use of fivefold cross-validation. The proposed method achieves a high level of accuracy, which demonstrates its usefulness for disease management and precision agriculture.