NIRDAM PROJECT

Near Infra-Red Drone Agricultural Monitoring

The objective of this project is to create an affordable precision farming and field-monitoring technique. We offer a solution that reduces agricultural risks and improves yield.

The technology developed within the project will enable the use of inexpensive precision-agriculture microclimate monitoring stations and UAVs to achieve monitoring of parameters relevant for agricultural production, previously available only through the purchase of expensive satellite imagery.

Our aim is to use NIR spectroscopy techniques with the combination of UAVs and scan the fields. Upon a successful flight, the data will be processed by software which will give reliable and accurate information about the soil properties.

Our data gathering activity (which will last throughout the entire project development) will consist of scanning different types of soils in the Vojvodina region with the help of precision-agriculture microclimate-monitoring stations in order to acquire a continuous stream of “ground truth” data. Vojvodina is particularly suitable for data gathering, since it has such a large diversity of soil on which farmers grow their crops such as: cambisol, albeluvisol, chernozem, terra rossa and fluvisol to name a few.

PanonIT - NIRDAM PROJECT

Enhancing Agriculture with UAV and ML

PanonIT - NIRDAM PROJECT

We plan to build precision-agriculture microclimate-monitoring stations (equipped with state of the art sensors), which will be deployed on all different soil types found in Vojvodina, in order to monitor soil properties continuously. In parallel with these measurements we will do aerial NIR spectroscopy with the help of UAVs. We will then attempt to use machine learning to build NIR-image-based soil-type-specific estimators for different soil properties monitored by the microclimate stations.

The models developed will then be used to derive precise soil properties information for the whole fly-over range of the UAV. Both reference-station independent and solution requiring the positioning of one or several stations in the fly-over range will be considered.

The arable land properties we will focus on with drone will be: Leaf wetness, solar radiation, PH level. Furthermore, we will attempt to use machine learning to develop a decision support system based on UAV-acquired data that will be able to accurately predict: need for irrigation, need for Fertilization, plant health.