Smart soil quality assessment for data driven fertigation using internet of things and deep learning

Abstract

Internet of things (IoT) in conjunction with artificial intelligence techniques are increasingly being used in different sectors, particularly in agriculture it is used to satisfy the need of increasing farm productivity to meet the rapidly growing demand for food due to rapid population growth. Ensuring data driven solutions is an essential step toward increasing productivity while at the same time enhancing utilization of resources. There is a need to ensure efficient usage of both water and fertilizer to conserve the environment and reduce costs. Even though attempts have been made to come up with smart irrigation and fertilization solutions, there is still a need to incorporate latest sensing and data analytic technologies. This study therefore presents a prototype for an Internet of Things (IoT) and a deep learning driven solution for smart fertigation. Soil moisture and soil nutrient data are collected in real time by sensors in a farm, data is then processed on an Advanced RISC Machine (ARM) Cortex 4 based Arduino Nano 33 BLE sense and a deep learning model used to predict when to irrigate. In case deficiencies in soil nutrients are detected, an alert is sent to the farmer via Global System for Mobile communication (GSM) module to add the needed fertilizer to the water. The irrigation valve is automatically actuated based on the predictions and from time to time. Upon successful implementation, this system will reduce water and fertilizer wastages while increasing productivity.