Supervised Learning System for Pest Disease Identification and Recommendations using Wireless Sensor Network in Agriculture Domain

Abstract

Designing wireless sensor networks (WSNs) that will sustain for longer time have been a challenging issue. For achieving this, efficient utilization of available energy should be anticipated from energy-constrained sensors running autonomously for long periods. In the field of agricultural research, farming related activities, assistance to farmers using advanced IoT and sensor network technologies is a need of time. Initially this paper briefed about smartphone based application systems that forewarns the farmer about pests/diseases that can occur on the crop. The designed system assists farmers to foresee the crop disease using machine learning algorithm. The disease can be predicted based on the environmental parameters as soil moisture, temperature and humidity obtained from sensor node and other parameters like rain fall, evaporation rate and sun shine hour obtained from open weather map application programming interfaces. A prototype is developed using ultra low powered micro-controller with interconnected sensor’s module. While comparing with the Arduino micro-controller based system, it was observed that the developed prototype system utilized less energy with effectively forewarned the farmers about crop diseases using Internet of Things (IoT). Thus it would assist the farmers proactively to reduce the losses due to crop diseases.