Detection of Potato Diseases using Deep Learning

Abstract

Potato is one of the essential foods and root vegetables in the world and is also the staple food in some different nations. Globally It is the fourth-largest food and now the seventhlargest growing food in Bangladesh. Most of the time, the fungus infects potatoes. Our farmers' profit, food demand, and food security majorly depend on Identifying proper diseases and giving solutions. At present, Image processing technology mostly performing the detection of identifying diseases. So, we used six different architectures of convolutional neural networks (CNN) in the deep learning field. The architectures are MobileNet, Xecption, Inception V3, VGG16, ResNet50, VGG19. From these architectures, one of the most popular CNN architecture is MobileNet that found the best result in our observation. For this approach, we used the potato image dataset. Here classified our dataset into three different classes where infected potatoes are two classes and a fresh potato class. This work can help to detect the disease of potatoes. This system will make it easier for farmers and researchers to find out potato diseases. It is much easier to Identify potato diseases in this method than to detect and classify the potato diseases manually, and it is possible to Identify potato diseases in a short time. This will be helpful for those who will be working with CNN.