Mustard Downy Mildew Disease Severity Detection using Deep Learning Model
Rishabh Sharma, Vinay Kukreja, Sakshi Sakshi
Abstract
Plant crop disease detection is becoming a more active topic of research as the requirement for these systems and approaches grows, as crop diseases have become a prevalent aspect of agriculture. In response to this requirement, we proposed a multi-classification model based on the deep learning (DL) approach. A Convolutional neural network (CNN) based multi -classification model that categorizes 1500 collected images of the mustard crop with healthy and mustard downy mildew (MDM) infected images based on MDM disease severity degree, as well as a binary classification to simply classify them. There were four different illness severity levels considered in the paper. For binary classification the accuracy of 95.6% has been achieved and for multi -classification, achieved accuracy is 96.66%. This research will contribute significantly to the field of using DL approaches to identify and detect mustard illness.