Plant Leaf Disease Identification and Classification Using Ensemble Learning and Mish as an Activation Function

The diagnosis and classification of plant leaf diseases in their first stages are crucial for enhancing crop output and ensuring food security. This paper presents a strong model for plant disease especially for the leaves through classification using deep learning the ensemble methods together with Mish activation function for accuracy and performance. The feature extraction is done by applying ResNet50, a deep CNN, and ensemble classifier techniques such as AdaBoost, bagging, and both hard and soft voting to improve the classification. The incorporation of the Mish activation function enhances learning by improving convergence and generalization of the framework. The model is implemented with TensorFlow and Keras to support deep learning and an application built on React Native that enables disease identification on mobile devices in real time. The authors adopted a proposed approach in classifying the plant diseases and the results show that it obtained a 98.3% accuracy. This solution, adaptable across several platforms, offers effective and economical methods for farmers and agricultural specialists to sustain plant health and avert crop losses. The presented model and application were tested on a wide dataset, and the results observed can point to possible use of the proposed method in precision forming.