Dynamically processing agricultural data from controlled legume sites

Smart agriculture is becoming a discipline with increased data processing requirements due to the need to understand complex agricultural ecosystems. In these ecosystems, modern digital technologies, that extend conventional tools (e.g., rain gauge, thermometer etc.), are being used to continuously monitor the physical environment resulting into large quantities of datasets in an unprecedented pace. In order to get value out of this data, efforts are being harnessed to process the data and use the outcome to make decisions that affect farm management tasks. Most of the existing data processing approaches focus on batch processing with periodic feedback. Even though this is useful, it is more desirable to have the data processed in a timely manner to support farm management decisions. In this paper, we propose our vision for a dynamic data processing platform for smart agriculture called DyPro. The platform allows processing of data as soon as it arrives to the backend and feedback sent to control points for action. To validate our platform, we present a scenario where data is being emitted by sensors into a central server for processing as soon as it arrives, and feedback sent to actuate required measures and controls.