Classification of potato diseases using deep learning apprach

Non-destructive and rapid inspection utilising deep learning and image processing may be of tremendous use for potatoes and other agricultural products. In this study, we recommend employing a deep neural network to separate diseased potatoes into groups according to their outer appearance. Train, test, and evaluate the resulting DNN using a collection of potato pictures. Using Machine Learning, a Modified Support Vector Machine (MSVM) and a Conv. Neural Network (CNN) were used to make predictions and comparisons about the segmentation and feature extraction of a potato picture. Canada and the United States Department of Agriculture standard datasets were modified for the purpose of testing the suggested model. The results show that deep learning is superior to both traditional methods and the Machine learning model. In some sort of picture, more than 99 % accuracy recorded, while in others, approach 97 percent.