Convergence of machine learning, deep learning, and IoT for plant disease detection

Detection of pests in the crops and their control at the right time are now very important activities in agriculture. In an agriculture-based nation, losses in agriculture not only affect the farmers but also affect the economy of the country to a greater extent. With the growing rate of population, there is a parallel increase in demand for food in the world, which is becoming one of the most important concerns for the developing as well as developed nations to fulfill the demand of their country. India, being an agriculture-based economy, faces a lot of challenges in not only producing ample amounts of food but also in ensuring the quality of the crops. It has been observed in the past few years that there is no certainty of a good yield of crops, sometimes because of weather and sometimes because of pests. Thus, detection of pests in the crops at the earliest stage has become an issue of concern. Many efforts have been made to detect and stop pests in agricultural lands at the initial stages. Manual monitoring of crops in the fields is not an effective way of controlling pests. In this phase of emerging technologies, it is a better way to tackle the problem of pest detection and control with the help of modern technology. A lot of solutions rely on sensor (image sensors and acoustic sensors)-based detections. The data collected through sensors are fed into deep learning algorithms and inferences are made, but because of many physical conditions and limitations, it is difficult to get real-time data and detect them. This article discusses various methods that are applied to detect pests in the crops. The aim of this article is to describe some of the most relevant research about pest detection using image processing and machine learning techniques; to give detailed reviews about the research studies carried out so far, with a special focus on the drawbacks of these technologies and future scope; and to provide some workable solutions to existing problems by measuring the biological response of the plant when it is affected by the insect. Measurement of DNA and RNA content of plants can work out as an efficient solution in the early detection of pests.