Employing a hybrid lion-firefly algorithm for recognition and classification of olive leaf disease in Saudi Arabia

Abstract

For the proper management of Saudi Arabia's olive harvests, a key agricultural resource in the area, the prompt and accurate detection of Olive leaf diseases is of utmost importance. Due to Saudi Arabia's varied environment, several diseases that harm olive leaves have emerged. Additionally, variables associated with climate change have made the spread of illnesses like the anthrax infection harsher. Although several methods for forecasting olive leaf diseases have been created, overfitting, categorization accuracy, and detection effectiveness remain problems. The Hybrid Lion-Firefly Optimization (HL-FO) method, a hybrid algorithm that incorporates the Lion and Firefly algorithms, is the innovative solution proposed in this study. This hybrid approach aims to improve the precision and effectiveness of feature selection and categorization approaches in identifying and categorizing olive leaf diseases in Saudi Arabia. The suggested system incorporates a Convolutional Neural Network (CNN) to categorize the picture and reduce overfitting problems. CNNs are a viable method for properly classifying various olive leaf diseases since they have been demonstrated to be very efficient in image classification tasks. Receiver Operating Characteristic (ROC) curve analysis, the F1-score, recall, accuracy, and precision are used to systematically examine the performance of the suggested hybrid technique.

Keywords

• Olive leaf disease
• Hybrid Lion-Firefly Optimisation (HL-FO)
• PCA
• Deep Learning (DL)
• Grey Level Co-occurrence Matrix (GLCM)