Early Stage Prediction of Plant Leaf Diseases Using Deep Learning Models

Abstract

Plant diseases lead to yield losses of up to 30% by each year, resulting in huge financial losses for farmers as well as a threat to world food security. As a result, early detection of infections in plants is crucial. Detecting infections in plants have always relied on human explanation via visual assessment. The technological advancements in the field of image processing have now made it feasible to give a variety of image-based analysis at a significant level. In the automatic disease recognition of various plant images, machine learning methodologies were often applied, majority resulting in low precision and defective extension. Utilizing deep learning in disease identification made it possible to improve the classification accuracy. In the field of image classification, the most recent generation of convolutional neural networks has shown notable results. The proposed XGBoost with CNN-SVM classifier is shown to be a fast, extremely efficient method for classifying specific imaging features into desired disease classes, as well as giving preferable results over the plain CNN and other classifiers, such as the support vector machine (SVM) for large datasets. Finally, the experimental findings show that the suggested system achieves a maximum classification accuracy of 97%.