Systematic Review of the Early Detection and Classification of Plant Diseases Using Deep Learning

Abstract

Abstract Deep learning (DL) addresses the brilliant period of Artificial intelligence (AI) and is slowly developing into the main technique in numerous fields. Currently it assumes a significant part in the early location and order of plant diseases. Plant diseases have long been one of the main threats to food security, significantly reducing crop yields and quality. Therefore accurate disease diagnosis is the main goal. The utilization of machine learning (ML) innovation in this space is accepted to have prompted a huge expansion in usefulness in the hydroponics area, particularly in the new rise of ML which appears to expand the degree of precision. As the latest modern technology in image processing and successful application in various fields, deep learning has great potential and broad prospects in agriculture. This paper surveys 40 studies using deep learning methods applied to agriculture and food production. In this study, deep learning is compared to other popular image processing techniques. The findings show that deep learning provides better performance. Future directions may additionally consist of the usage of drones and agricultural robots to automate photo seize and then zooming in on plant sickness image datasets, using newly posted fashions that describe more efficient architectures with fewer parameters, as well as the use of new techniques for photograph enlargement inclusive of generative adversarial networks (GANs).