Plant Disease Detection and Classification Using Artificial Intelligence Approach

Abstract

Agricultural development is not only crucial for human existence; it is also a major commercial field for many countries, regardless of their state of development. Early detection of plant diseases is crucial because they have an impact on the growth of the affected species. However, plant diseases have become more prevalent recently for a number of natural and artificial reasons. Globalization, trade, and climate change have all had an impact, as has the deterioration of existing systems as a result of years of agricultural intensification. For a long time, plant disease detection using computer vision has been a great topic of discussion. Emerging technologies such as artificial intelligence can be used on images of plants to detect infection and sickness in plants, which will allow people to take required steps which can heal the plants. As the datasets on this work are restricted, disease identification is mainly done through images. The goal of this work is to accurately diagnose sickness in plants (among a few possible diseases present in the dataset) by comparing and selecting the model with the highest accuracy on training and testing data. Common datasets of plants with their disease are introduced, and the results of previous research are compared. On that premise, this work examines potential problems in practical implementations of deep learning-based plant disease diagnosis. This analysis also provides aspects to focus on in order to improve this project. Our research also looks at how machine learning approaches have evolved in the last few years, from traditional machine learning to deep learning.