Explainable AI for Crop disease detection
S Rakesh, M. Indiramma
Abstract
Agriculture is a major contributor to the Nation’s economy since it is the most important component. It is common for plants to suffer from various kinds of diseases, which may be influenced by extreme climatic conditions and pests. This further lowers the quality of the harvest. In order to calculate the overall crop yield and productivity, early detection and accurate diagnosis of plant diseases are critical, which can significantly increase productivity and yield. Farmers have difficulty identifying diseases in plant leaves. Diagnosing diseases through traditional methods requires extensive field experience & expertise. The advancement of technology has allowed numerous strategies to be developed to identify plant diseases using Artificial Intelligence and Deep Learning. To detect and classify various crop diseases, we propose two models, Inception V3 and ResNet-9, which are deployed on Plant Village Dataset and New Plant Disease Dataset. In addition to Deep Learning models, Explainable AI (XAI) tools such as LIME and Grad-CAM have been utilized for understanding the black-box nature of Deep Learning models.