An Empirical Survey of Machine Learning-Based Plant Disease Prediction Models

Abstract: The occurrence of diseases in plants badly impacts the agricultural production, which increases the food insecurity when the diseases are left undetected. Particularly important for ensuring the availability of production of agricultural and food are the major crops, such as maize, rice, and others. Effective control and prevention of diseases in plants are based on disease forecasting and early warning, which is essential for managing and making decisions regarding agricultural productivity. In rural parts of developing nations, observations by knowledgeable providers remain the main method for plant disease identification as of yet. This draws researchers in for ongoing experienced monitoring, which may be cost-prohibitive on large farms. Besides, in some remote areas, farmers require the assistance of the agricultural experts, which is the expensive and time-consuming process. Hence, automatic disease identification for plants is important to promote the monitoring of large crop fields, which encourages the contribution of the accurate, less-expensive, automatic, and fast technique to perform the detection of diseases in plants. In this survey, the automatic detection methods used for the plant disease detection based on the deep learning methods are discussed. The importance of the deep learning methods for the detection of disease is demonstrated through the schematic sketch on the other basic machine learning techniques in agricultural applications.