A Smart Wireless System to Automate Production of Crops and Stop Intrusion Using Deep Learning

Abstract

The idea of automating the production of crops has existed since the early 90s and one of the major issues both scientists and farmers face, is the question of irrigation. An irrigation system is a dynamic system that is predominantly dependent on external covariant. This paper provides a methodology by utilizing a custom-built mathematical model which includes wireless sensors as a data source that is processed on Google Cloud there by providing a smart IoT enabled architecture that can be scaled even on large farms. Based on Holistic Agricultural surveys, around 35% of crops get destroyed due to animals and humans. This intelligent system is enabled with deep learning neural networks implemented with Tensorflow to identify animals based on its threat level and human intruders who are not authorized on the farm, and immediately report the intrusion to the farmer. The system is equipped with an android application that provides remote access to the system and surveillance through live video streaming.