Cloud-Based Analytics for Sustainable Agriculture: Leveraging AI to Bridge Farming and Rural Health Outcomes

Abstract

The global population is projected to reach over 9.5 billion and possibly up to 10 billion by 2050. Its growing economy requires increases in sustainable food, fuel, feed, and fiber for global security. The agricultural sector faces grand challenges to meet these increased demands under the constraints of climate change and dwindling natural resources with limited arable land and fresh water. Combination of intensifying production and expanding crops has led to serious challenges such as worsening soil and water quality, greenhouse gas emissions, and crop productivity sustainability. Sustainability of agriculture under the strain of grand challenges depends on coalescing affordable and reliable sensors and IoT instrumentation, advanced computing power and algorithms in data processing and machine learning modeling, and secure internet connections with portable and user-friendly interfaces and user experiences. The rapid adoption of IoT data sensing technologies in agricultural settings brings new opportunities to help bridge farming practice and rural health outcomes at both behavioral and clinical levels. While sensing technologies are arguably more affordable, accessible, and versatile than ever, the sheer amount of data is overwhelming.
In agriculture, farm management systems and other platforms have been providing various forms of decision support for on-farm data collection and analysis. While increased data innovations, availability, and digitization are advantageous, they bring data inflation challenges and data-related issues to agricultural producers. Data generated from different disciplines can be highly heterogeneous. Datasets across disciplines may not share the same ontology, modality, or format. The growing amount of data diversity presents additional challenges. If left untamed, it may lead to underutilization of data information and opportunity, mistaken insights, and degraded trust. Agricultural data, such as remotely-sensed satellite, aerial, drone, weather station data, and on-site soil, elevation, land use, pest, and irrigation data, are often big and complex. The data types are highly structured and may contain both temporal and spatial dimensions. First, data is collected by various types of telemetry systems and machine-level devices. Then, data is exchanged and transmitted from cloud-based systems for data fusion and cross-scale information extraction. Standard data sharing protocols are needed to ensure the cross-compiling capability of data services and applications such as remote sensing data safety and security.