Applying a Deep Learning Approach for Wheat Rust Disease Detection

Abstract

Crop reduction is the major problem the world is facing currently in ensuring food security and creating a sustainable agricultural system because of different crop diseases. Tracking the health status of crops is a critical job we have to do to prevent the spread of different crop diseases and it should be in a technological manner rather than by the labor force. This study presents developing a model that detects three types of wheat rust disease detection using a Deep Learning approach, especially the Convolutional Neural Network (CNN) by using the image of the wheat crop. These wheat diseases are wheat leaf rust, stem rust, and yellow rust which also differ in the way they affect the crop and in their level of damage. Color code segmentation is proposed in this study to extract only the needed information from the wheat image to identify the healthy crop from the infected one. After conducting more than 200 hundred experiments using different impacting factors such as Learning rate, Dropout, and train test split ratio the study has finally achieved a 99.76% accuracy to detect the wheat rust from healthy crops