Enhancing Maize Leaf Disease Detection Using Cat Swarm Optimization with Deep Learning Model

Artificial Intelligence (AI) has been used for various applications including communication, medical, object tracking, and object detection. The major crop in the world, Maize crop is affected by various disease kinds which could affect the quality and lower its yield. Classical Machine Learning (ML) approaches, such as classification and feature extraction, make extensive use of plant disease detection. These techniques can extract features from image datasets, like shape, color, and texture, for training a classifier that could distinguish between diseased and healthy plants. Deep learning (DL) and AI-based approaches are increasingly used in agricultural research due to their capability in learning the deep features automatically from the images and also their speed and accuracy levels are increased than the classical methods. This study proposes a novel Maize Leaf Disease Detection employing Cat Swarm Optimization with DL (MLDD-CSODL) method. The proposed technique includes different phases, beginning with contrast enhancement to enhance the input image quality. The feature extraction stage applies Capsule Network, a DL algorithm known for its capability to capture hierarchical relationships amongst features. To improve the model's hyperparameter, Cat Swarm Optimization (CSO) is used, enhancing the overall performance and efficiency. The last classification stage applies XGBoost, a robust ML classifier, to correctly categorize and identify maize leaf diseases according to the features extracted. This wide-ranging MLDD-CSODL algorithm aims to increase the speed and precision of disease detection in maize crops, providing a potential solution for earlier intervention and better agricultural management.