Apple Fruit leaf Disease Detection and Classification using Quantum Behaved Particle Swarm Optimization

Detecting and classifying apple fruit leaf diseases manually is a time-consuming and expensive task, requiring expertise. This process not only affects the quality of the leaf disease image but also demands expert intervention. In the dataset, there are 1615 healthy images distributed among four classes: apple scab, apple rust, apple black rot, and healthy apple. Data augmentation techniques were involved image inversion, noise injection, colour enhancement, rotation and scaling. Despite these efforts, the existing model faces challenges in an inefficient search space with high dimensions and local optima. To enhance the Quantum-Behaved Particle Swarm Optimization (QBPSO) based Long Term-Short Memory (LSTM) dimension the search space is expanded globally, independently detecting the leaf disease. Feature extraction is performed using Squeeze Net layers to analyse the leaf disease, reducing parameters to improve the detection process. The fire modulus in the analysis is involved in the process. The hyperparameter tuning process utilizes QBPSO to independently detect and classify the leaf disease, with the Schrödinger equation playing a crucial role in the process. The obtained results demonstrate that the proposed method achieves better accuracy of 98.76%, Precision of 98.72%, Sensitivity of 98.82% and F1-score of 98.70%. on the plant dataset. These results ensure detection and classification performance compared to other existing methods, such as Discrete Wavelet Transformation (DWT), Butterfly Optimization Algorithm (BOA) and Chaotic Salp Swarm Algorithm (CSSA).