Cotton Leaf Disease Classification using Deep Convolution Neural Network with Explainable AI
Samarth Otiya, Pavan Faldu, Parth Goel
Abstract
The cotton leaf disease detection model combines Convolutional Neural Networks (CNN) with Explainable artificial intelligence (XAI) technology for the goal of diagnosing diseases in cotton leaves. This diagnostic procedure incorporates the classifying of leaves into four unique classes: healthy, curl virus-infected, bacterial blight-affected, and fusarium wilt afflicted. In contemporary times, agriculture and farming have witnessed a significant evolution, with disease detection techniques playing a pivotal role in this transformation. This technology proves invaluable to farmers by facilitating to detection of diseases in crops after that farmer can empowering them to take proactive measures to manage and mitigate the impact of these diseases on their agricultural yields and overall crop health. The primary objective of this research endeavor is to classify diseases within a set of four distinct categories. This categorization challenge is done by the employment of Convolutional Neural Networks (CNN). The research leverages transfer learning models, including Xception, ResNet50, DenseNet121, and EfficientNet, to achieve high classification accuracy while also employing Grad-CAM for disease detection. The projected accuracies for each model are as follows: 85% for Xception, 99.1% for ResNet50, 93.3% for DenseNet121, and 98.5% for EfficientNet.