Land Monitoring and Sustainable Management

This pilot aims at providing farmers and decision makers with real-time with actionable intelligence upon which to monitor soil health, increase agricultural yields, optimize resource consumption and sustainably manage land for a better future. The idea is to obtain an online framework that captures and exploits the unique data from the hyperspectral and soil-condition sensors in combination with EO data. This data will improve the existing Open Land Use database. A semantic data repository will provide access to a two layered data model. The first layer will be based on open-access Level-1C multi-spectral imaging products from the Sentinel-2. Sensor data from the field robots will be fused to enrich this base layer (including UAV-enabled hyperspectral imagery). This data layer will also provide long-term data preservation abilities, in preparation for wide-scale deployment and future big data applications.

With a fast query interface, algorithms to request and exploit the data will be developed through deep-learning algorithms to provide agromachinery and robots, precision targeting and intervention capabilities by interpreting sensor inputs (e.g. plant health, nutritional status, weed detection, soil health) and their correlation to ground truth. Learning algorithms will be trained to infer at the lowest level, gross areas that require proximal on-site intervention from EO and UAV hyperspectral imaging. By creating a hierarchical, closed loop learning process

through multi-resolution and multi-band field sensing we will enable knowledge-guided and precise farming interventions but also create a framework for future work in multi-site data mining and learning (at farm, regional and cross-border levels). EUXDAT e-Infrastructure will provide the means to facilitate data connection (including in realtime) and moving to the computation targets, where data will be processed in an efficient way. The End User platform will facilitate the definition of the tools and algorithms to involve in the workflow to be followed for completing the solution, which will be accessed in a transparent way for the users.

Energy Efficiency Analysis

The energy sector is experiencing an increasing tension between safety and cost-saving requirements. Of late, the need to ensure Europe’s long-term strategic interests became another factor contributing to this tension. In this context, agricultural production influences, and is influenced by, water quality and quantity, ecosystems, biodiversity, economic activity and energy use and supply. The seasonality and ubiquity of agriculture make agricultural practices and production amenable to efficient synoptic monitoring. The effectiveness of production, including agriculture, is determined by the ratio of the value of the production outputs to the value of production inputs. When it comes to agricultural production, the efficiency is affected not only by the internal factors of the production process, but also by external factors such as climate, subsidies, the situation in the global market, among others. Furthermore, acquiring knowledge about the energy and carbon intensity of different crops on different lands, about the way farm processes work and how to take care of the variability within fields in a single farm is very demanding task. New technologies are needed to collect sensitive data and evaluate it in the most accurate way. No optimization process can be performed without sufficient and objective knowledge. The pilot will focus on estimating of energy efficiency of the production process.

We will use exploratory visualisation methods, combining different types of data like yield potential from Copernicus, DMT, Machinery monitoring data, Land Use data, Road Network, historical data to analyse processes, cost and benefits on level of every field e.g. through an algorithm for optimal trajectory of farm machinery, using the EUXDAT e-Infrastructure tools. Exploratory visualization is the process that involves an expert creating maps and other graphics while dealing with relatively unknown geographic data. Generally, these maps serve a single purpose and function as an expedient in the expert’s attempt to solve a particular (geo) problem. While working with the data, the expert should be able to rely on cartographic expertise to be able view data from different perspectives. As such, the resulting maps and graphics are available in an interactive viewing environment that stimulates visual thinking and promotes informed decision-making using WebGLayer. The technology is currently used for Farm Telemetry tasks and for the analysis of Yield potential.

3D Farming

3D precision farming is a new approach and will help to better plan and manage a farm production. Current practically utilized precision farming systems are working only with 2D data analysis and 2D visualization. On the contrary, e.g. soil erosion or nutrients balance are significantly influenced by the vertical aspect. There are new approaches focusing on 3D, using different types of DEM including UAV. However no one incorporate usage of the 3D visualization and analysis of existing data, even if it can explain many problems inside the field. It all leads to soil erosion prevention, reduces nutrient leaching and therefore to cost effective land management.

The pilot will build mainly on: Copernicus Sentinel 2 data, potentially also Sentinel 1 and Copernicus Digital Elevation model, Landsat 8 data and available indexes (NDVI, LAI) and yield potential derived from this data Open Land Use Data, Land Parcel Information System data, Soil Maps and National Digital Elevation Model as well as machinery monitoring data. We will also include historical meteorological data. Usage of meteorological data in combination with yield maps will help us to better understand dependencies among processes. The focus will be on next tasks: Visualization of the land use and soil productivity data together with DEM derivate (elevation, slope, aspect) and meteorological data.

The pilot will perform spatial analysis searching for the highest productivity land/zones (with appropriate humidity, rich to fertilizers, etc.) in well accessible areas (e.g. small slope, +/- south aspect) using EUXDAT connectors, data management tools and data analytics tools. We will also perform 3D visualization of Yield potential Maps and Open Land Use maps. The combination of this data sources will bring new research methods into PF. The usage of 3D analysis will help to better understand conditions on the fields and will help to reduce amount of fertilisers. It is known, that elevation has a large influence distribution on water, nutrients and soil particles in the field. This will help to reduce utilization of chemicals and decrease costs of farm production. Usage of 3D modeling for planning of operation, machinery movement and crop rotation will have large influence on erosion and soil protection.