Bangladeshi Local Potatoes Dataset and Classification Using Deep Learning

Abstract

Potato is one of the major foods in Bangladesh during winter which shows the most environmentally amiable vegetables. Variations of potatoes in Bangladesh that are grown here are extensively categorized into two divisions, high yielding and local. Assuredly, potato is the world’s most important crop when it comes to rank according to the volume of the fresh products. Moreover, the minerals, fibers, and vitamins it delivers can serve convenience for human healthiness and guard against diseases. Recently, different kinds of potatoes of Bangladesh has been exporting to various countries worldwide. Hence, a lot of people are engaging in agriculture to cultivate potatoes. It is very essential for the farmers to acknowledge that; which kinds of potato cultivation will be beneficial according to the market price and cultivating process. This system can help agriculturists, farmers, and general people to find out the local species of potatoes. This dataset of locally recognized potatoes like Diamond Alu, Mete Alu, Lal Alu, Jam Alu, Lal Mishti Alu, Chhoto Lal Alu, Mete Chhora Alu, Shunno Alu, Shada Mishti Alu, and Keshar Alu. Using a dataset of potatoes (5190 images) that have been collected by us, this research trained a CNN (Convolutional Neural Network) and deep learning based model to identify potatoes. The model gets an accuracy of 98.87%.