Plant Disease Detection Using Deep Learning and Machine Learning Algorithms

Understanding the economic sector of India's agricultural productivity relies heavily on the ability to identify diseases in plants. Although numerous diseases pose a significant threat to the food industry, prompt detection remains challenging due to a lack of infrastructure. Categorizing the diseases in plants at an early stage can help in the proper growth and development of the species. The increased use of computers has made deep learning possible and convenient. This research paper explores the use of deep learning algorithms, specifically ResNet model for the automated classification of plant diseases found in the images and many more have been explored to detect the diseases in plants using a dataset of plant leaves with diseases. The goal is to check how well the algorithm works and figure out which one is best for classifying plant diseases. This paper proposes a model which includes image filters and deep learning for plant diseases and it is found that ResNet is a powerful architecture for plant disease classification and achieves 87.5 % accuracy. Approaching the methods of deep learning on images of leaves shows a clear path towards the disease in the plant on a global level.

I. Introduction

Modern technologies in India have given the ability to humans for producing an ample amount of food for the population. But, the food industry is threatened because of factors like changes in climate, diseases, and others. Every culture on the planet relies heavily on agriculture as a source of support. Agriculturalists have a huge variety to select the vegetables and food crops. The diseases in plants are not only the fear for the food industry but also for the farmers whose livelihood depends on how healthy are the crops. There are many practices namely crop rotation, pesticides, irrigation, etc. which are being performed for a very long time now. There are many approaches to agriculture but proper management is needed for it to survive. In the agriculture industry, diseases in plants and the use of fertilizers are the main reasons for the heavy losses of crops. Detecting the diseases at an early stage is appreciated as at the later stage it can affect the animals and the human beings. Diseases in corn, tomatoes, potato, and other plants occur in different parts like the root, stem, leaves, etc. To classify images accurately large data sets are needed, of both diseased and healthy plants which are verified. Major diseases are apple scabs, black rot, and cedar apple rust which are affecting in apples. Diseases that are affecting potatoes are an early blight and late blight. In the case of corn (maize), diseases are grey leaf spot, common rust, and northern leaf blight; diseases in grapes can be black rot, leaf blight, and esca; in the case of pepper, diseases are bacterial, whereas tomato leaves can have target spot, leaf mold, mosaic virus, yellow leaf curl virus, bacterial spot etc. The diseases of tomato, potato, apple, corn (maize), grape, and pepper bell are shown in Fig. 1.