A Crop Water Stress Monitoring System Utilising a Hybrid e-Infrastructure

Abstract

There are many challenges involved in irrigation systems in agricultural environments. Many approaches simply support crop watering as a simple standardised solution - watering the fields uniformly. However, depending on the topology of the ground and the crop species that exists, considerable improvements can be achieved that can greatly impact on the overall crop yield. Too little watering or too much watering in the wrong areas can impact the overall output. With the rapid growth in unmanned aerial vehicles (UAVs) and drones, image capture technology is now readily available. If the water-stressed areas could be identified, irrigation systems could offer targeted irrigation to only those crops in need. To support this, a crop water stress monitoring system has been designed utilising the thermal images of the crops. The analysis of such data can be computationally demanding when large crop fields are considered, hence the system has been designed to use multi-core high performance computing and Cloud-based systems. This paper outlines the requirements and design of the crop water stress monitoring system and how it provides an efficient and effective analysis of crop water stress index. The paper also benchmarks the algorithms across the hybrid infrastructure.