Comparative analysis of different plant leaf disease classification and detection using CNN

Increasing threat to global food security, efficient and timely detection of plant diseases is crucial for ensuring agricultural productivity. The paper deals with a comparative analysis of the accuracy of the CNN applied for leaf disease detection in two major crops: maize and soybean. The preferred methodology utilized the ability of deep learning to improve the accuracy and speed in identifying diseases, thereby aiding farmers in making informed decisions for crop management. The study employs a dataset comprising annotated images of maize and soybean plant leaves affected by various diseases. A deep dive into the architectural nuances of CNNs is undertaken, with a focus on optimizing model performance for both crops. Transfer learning techniques are explored to harness pre-trained models, enhancing the network's ability to generalize across diverse disease patterns. The training process involves fine-tuning the CNN models on the plant-specific datasets, optimizing hyper parameters, and employing data augmentation strategies to address limited training samples. The comparative analysis extends to evaluating the model's performance specifications such as accuracy, precision, recall, and F1-score, to ascertain their productiveness in disease classification. Additionally, the research investigates the robustness of the models against environmental variations. The authors explore the impact of model interpretability on decision support for farmers by visualizing the areas of interest within the leaf images that contribute to disease prediction. The results reveal insights into the relative strengths and weaknesses of the CNN models for maize and soybean disease classification, offering valuable guidance for stakeholders in precision agriculture.