A Neural Network-Based Approach for Pest Detection and Control in Modern Agriculture Using Internet of Things

1. Introduction

A system is required for farmers to foresee the crop requirements, so that they may enhance the crop quality and its quantity. Farmers have lack knowledge about current technologies that may help them enhance quality of seeds, irrigation facilities and good management of the crops. The delivery of crops in market gets delayed due to the lack of proper road connectivity between the rural and urban areas (Rashid, 2015). Also, climatic situations can become unpredictable hence resulting in wrong decisions by farmers. In mountain areas irrigation water is not easily available and cost much to farmers. Sensor networks are a solution to all these problems which may be utilised to determine the approximate values of field information. They also help in determining the simulation models of the crop, insects, diseases, and other data. Thus, farmers can monitor in real time the field conditions. How it is carried out will be discussed in next sections.

Integration of different geographical and temporal figure of the farm-land are easily achieved with the help of wireless sensor networks, also the variations in climate, pressure, soil humidity, motion, temperature, pest and suggesting best solution for management to the farmers are also possible with Wireless sensor networks.

The increment in wireless sensor network deployment in industrial, agricultural and habitat monitoring applications is due to the fact that this technology consumes low power and requires low data rates for transmission or reception of data, hence it proved to be an energy efficient technology. In addition to this it also provides mobility and flexibility in connection which enhance network expansion when required. The main objective of agriculture through WSN is to make crop production better by managing and monitoring the growth duration.

1.1 Introduction to Wireless Sensor Network

Each node equipped in a wireless sensor network includes a sensor module, a processing data storage module, a power supply module and a wireless transmission unit (Rashid, 2015; Navin, 2013). Every node is provided with the capacity that they are easily able to collect and process physical information so that the gathered data can be further sent to the monitoring station (base station) or to the sink node. Wireless sensor network consists of implementation of more than one sink nodes and several sensor nodes in a physical environment. The topology arrangement of Wireless sensor network nodes may differ from a star network to a multi-hop wireless mesh network. The processing technique between the hops of the WSN can be done by using either flooding or routing. One of the best features of these sensor nodes is that they can switch between various modes when required; the modes are sleep, active and idle. When not in use they switch to sleep mode to save energy, this increases the lifetime of the nodes and they can be used for a long period without replacement.

Figure 1.

Deployment of Sensor Nodes and their communication with the Base Station

Environmental information plays a very crucial role in the field of agriculture. This is because the crop production is mainly bank on the environmental situations and the feedback of the plant development to varying environmental situations is highly problematic (Holt, 2007; William, 2005). Farmers and researchers should collect the information at their areas on their own but the traditional climate stations are very expensive and much big in size. Till now, data loggers were used for this task but with a condition that the users should have to go at these stations frequently to gather the related data. Hence to solve all such related problems sensor networks are desirable.