Novel and Efficient Plant Disease Prediction Using ViT and YOLOv8 Deep Learning Models
Abstract
Plant health is affected by several abnormal conditions which indicate virus, bacteria and fungus caused by pathogens. Environment conditions cause plant disease such as water stress, pollution and nutrient lack. These diseases cause several symptoms on plants such as spots, stunted growth and wilting. Sanitation is used for curing plants from plant diseases. Plant diseases need to be monitored frequently and early disease prediction can save plant health. Image processing techniques used earlier for plant disease prediction and machine learning techniques replace image processing techniques in disease prediction. Deep learning models are introduced and show accurate results over machine learning. Proposed research uses YOLOv8, Convolution Neural Network(CNN) and Vision Transformer(ViT) deep learning models to predict plant diseases. YOLOv8 deep learning model predicts plant diseases with accuracy 98% over lower accuracy results from CNN and ViT. Proposed research is useful for agriculture field where plant diseases can be predicted more accurate than other existing works.