Enablers Marketplace

The DSS on animal welfare assesses the health and well-being of cows by analyzing various factors like nutrition, hygiene, rest, and movement, which are closely linked to their productivity. Knowage collects and processes data on milk production, cow activities, and rest patterns to display parameter values and provide insights into the cows’ health status.

A ML module employs the Random Forest algorithm to estimate traceability, specifically predicting discrete class labels like High, Medium, or Low milk quality. To train the model effectively, historical data relevant to the problem domain is required, allowing the model to learn the relationships between data and desired predictions.

Translation services from CSV to json-ld AIM for AnimalWelfare and MilkQuality.

The DSS for Nitrogen Balance Model utilizes NDVI images, crop coordinates, field properties, and meteorological data to estimate crop nitrogen requirements and create fertilization schedules. It generates cartographic plans and recommendations for farmers to implement differentiated fertilization within the same plot, aiming for higher yields.

To diagnose plant stress, the NDVI vegetation map composition is employed. Component offers REST APIs for sending input JsonLD files in AIM format via HTTP POST requests and retrieving algorithmic assessments of plant stress levels through HTTP GET requests. Internally, it accesses the NDVI Classification service to obtain satellite images from Terrascope and classify pixels based on color.

This component will expose a function via rest services and will act as a switch for the 2 blockchain present in DEMETER.

This component will expose 2 functions via rest services: Posting a JSON with the following format: {“Hash”:data} will store the data in the blockchain and return the transaction ID. Posting a JSON with the following format: {“txId”:transaction_id} will return the data.

Enabler Analytics for Maize Irrigation uses Machine Learning algorithms, and more specifically RNN that can provide forecasting for meteo data, and the user can decide if and where to irrigate maize plants or irrigate before and after the specified irrigation timetable.

The Milk Yield Prediction Algorithm is a pre-trained component specifically designed for the NRF breed. This algorithm is capable of predicting the remaining lactation yield curve for a cow based on its current lactation data. To make these predictions, it requires input data such as the animal’s identifier, daily milking weight yield, daily concentrate feed consumption, and the start date of the lactation (i.e., the last calving date).

Get realtime data of flaxing machines.

SeedsBit Traceability and Document Notarization Platform.

Soil Interpolation using soil moisture sensors and satellite imagery combined with machine learning model is innovative which provides soil moisture content at all points and depths of the farm. Such measurement and prediction of soil moisture can help growers on all scales. Informed knowledge of soil moisture allows users to target irrigation only in areas that need water.

This component enables data extraction through a query interface directly connected to the spatio-temporal semantic data management system of the AFarCloud smart farming platform. In turn, the component performs a runtime translation of the data for representation and integration into the Adaptive Visualisation Framework (AVF) platform. This service provides information on the observations.

PENFA DEE aims to demonstrate a highly innovative and interoperable system that will be able to integrate data from the ambient environment of greenhouses with the diagnostic data derived from BELD (BioELecticDiagnostics) device, regarding pesticide residues in critical growth stages, to enable and empower growers to perform optimal pesticide use in the critical fragment of greenhouse horticulture.

The pollination optimization service provides a unified communication mechanism (via programming API) for exchanging data between farm and apiary management systems. The service allows for matching farmers pollination needs with beekeepers offers and automating notification about planned sprayings. It allows for management and exchange of the beekeeping and farm information.

GEM4GF, green energy management for green farms developed an innovative solution for designing energy self-consumption communities for farmers. “Power farms” app, allows farmers to define elements of a self-consumption community such as farming houses, cooling facilities, irrigation for farmland as well as wind turbines and PV installations and compute the possibility for self-consumption (Austrian market relate model of energy communities.

This enabler makes it possible to visualize in “Knowage for DEMETER” the results of the data analysis. More than 15 machinery parameters ranges are checked and classified as good (green), problematic (orange) and highly problematic (red). It needs the enabler “Data Quality Assessment tool” for structured data. It returns an AIM conform file that can be imported into Knowage to get a visualization.

This enabler identifies driving sequences of machinery on the road, and performs a series of controls regarding oil temperature, heavy accelerations and break events. The AIM output file can be imported into Knowage for DEMETER for visualization and assessment of the driving (and thus emission).

This enabler is a subcomponent of the Data Quality Assessment enablers “DQA for structured data” and “DQA for linked data”. It accepts structured data: comma separated files, e.g., sensor measurements; JSON, e.g., metadata; and shapefiles, e.g., field boundaries or field application data. The output of the DQA is provided in machine-readable JSON format and can be used for further analysis.

The Data Quality Assessment (DQA) enabler can measure the data quality of linked data. The APIs of the DQA components can be used either directly within Python or can be deployed and used by external systems through a REST API. “Data Quality Assessment (DQA) Tool for structured data” is a separate DEMETER Resource.

This enabler provides interoperability of services through data integration. The component provides “translation” between observations modelled under the AIM JSON-LD, and the AFarCloud JSON data model. The component provides a REST endpoint with a swagger interface that, from an AIM-compliant input, generates its semantic container under the AFarCloud model.

Node-Express App transforming sensor information to Demeter’s AIM format.

The Data Transformation Enabler bridges the gap between the Agricultural Data Model (AIM) JSON-LD used in the DEMETER project and the AFarCloud JSON data model, enabling data interoperability. It validates, translates, and stores observations from sensors or devices in an optimized time-series database. This allows DEMETER to easily integrate AFarCloud platform services for data management and repositories.

This service provides information on the observations generated by static and mobile sensors and devices in a determined area. Thanks to the query interface, the management of large volumes of spatio-temporal data are enabled. Detailed documentation together with the service specification, connection points, and a deployment guide can be found in the file “DataQueryEnabler.pdf”.

This service allows the estimation of Water footprint (WF – Green Water and Blue Water) for the next 7 or 15 days (in AIM format), for a specific (AgriParcel) location of a “vineyard” field in Cyprus (CY), using data from a meteorological service (VisualCrossing). Therefore, the WF can provide an effective irrigation plan for a particular grapevine variety. Assessment of ET0 is done by means of the FAO Penman-Monteith equation.

This service allows to calculate Growing Degree Days (GDD) in AIM format, for a specific (AgriParcel) location using data from a meteorological service (VisualCrossing). GDD calculation is based on the “Averaging” method.

Component is developed for six fungal diseases and two insects that may occur in the vineyard and orchard, as a digital model, wrapped with an API and provided as a docker container image. The component takes environmental parameters into account to quantify if conditions for a disease and insect development are met and returns recommendations to prevent the disease and insects spread or to reduce its impact.

StreamHandler, an industrial-grade, Kafka-based message brokering tool. It offers full Kafka functionality in a secure, reliable, and scalable manner. This platform covers Brokerage Service Environment (BSE) pub-sub functionality. The platform can be also regarded as part of DEMETER’s data management system.

Knowage is an open-source suite which offers a wide range of features including data federation, data mining, and advanced data visualization. It is particularly tailored for big data analytics and supports rich and multi-source data analysis. The new Knowage BD product caters to various analytical needs within Big Data architectures.

This component retrieves a CSV file from cloud storage, processes it using a translator, and subsequently interacts with blockchain APIs to store the data within the blockchain. When requested by the user traceability dashboard, this component utilizes the transaction ID from the database to communicate with the blockchain APIs and provide the requested traceability phases’ data back to the dashboard.

This is a wrapper/translator that takes data (in csv format) from the ExMachina sensors located in the field and translates them to AIM json-ld format.

Component provides invaluable insights on the level of nitrogen based on satellite imagery, effectively supporting the farmers on whether, when and how much fertilizer should be applied. The output for every plot of land is a choice between 3 levels: “adequate”, “needs monitoring”, “immediate attention” and the results can be presented as a basic visualization over a field image in traffic light manner.

Component for Plot Agronomic Information. This component, exposes agronomic information of a plot, registered in the pilot cloud infrastructure, modeled using DEMETER AIM. To do so, it calls an entry point in the pilot cloud infrastructure to retrieve the data that the user wants/needs to expose.

This component provides an entry point to obtain the irrigation water estimation for a crop. It uses a mathematical model based on an update of the procedure for calculating the reference evapotranspiration and crop evapotranspiration from meteorological data and crop coefficients, which was presented by first time in the publication of the FAO Irrigation and Drainage Series.

This component will predict the reference evapotranspiration (ETo) value for the end of the day and next days.

Middleware connects LoRa Data packets from Chirpstack (HTTP integration) to Orion Context broker. Once an HTTP request coming from Chirpstack is received, the middleware is capable of detecting if a device is already registered in the Orion Context Broker. If the device is registered, the middleware updated its context information. If the device is new to Orion, the middleware registers it.

The DSS component for Irrigation Management provides the farmer with information to estimate the irrigation water needed for a crop using other components and data retrieved from the pilot cloud platform. This component will estimate the water irrigation needs of a crop using other DEMETER components: – Predicted evapotranspiration (ETo) – Plot agronomic info – Crop irrigation model based on ETo.

This enabler facilitates data exchange and interoperability between the components in DEMETER and the EU project AFARCLOUD. The enabler offers a REST service that takes as input an AFarCloud-compliant JSON file with observations from a multivariable sensor and converts it into a DEMETER-compliant JSON-LD file.

This component translates data from AIM ontology (JsonLD format) to a CSV data file containing the various vegetation indices. This is needed as input for the fertilization data analytics and optimization enabler in a CSV format. This wrapper takes AIM data and creates them in the form read by the enabler.

This is a simple GUI enabler that reads the fertilization needs (given as output by the fertilization enabler) and displays them on an image of the field.

This component provides capabilities for managing machine learning experiments and machine learning models, based on the open-source solution MLFlow. Most popular languages like python and R, and frameworks like scikit-learn and tensorflow are supported.

This component focuses on element identification in pictures. In order to do that, the component allows the creation of computer vision-based models from a zipped labelled images dataset and the use of those models in the identification of the elements labelled in the model generation dataset. As the component is AIM native, the input and the output are in AIM format.

This component will log the access to DEMETER resources by logging the issue and use of authentication and authorization tokens. These tokens contain the information about the user who is logged to the system and the resources the user is intended to access. The traceability agent will expose a REST API to register authentication and authorisation events (POST) and retrieve their details (GET).

Component generates a variable rate task map in AIM format for an agricultural parcel. It uses a WatchItGrow service to process Sentinel-2 NDVI images, taking into account field geometry, treatment details, and cloud-free image data. The resulting task map is divided into classes based on NDVI values, and these classes are translated into application rates relative to the specified base rate.

This component will provide access to the necessary functionality to predict the crop type for one or more agricultural parcels based on Sentinel-1 and Sentinel-2 timeseries. The component is equipped with a default LSTM classification model trained using reference data from Belgium and capable of predicting the several crop types (Maize, Wheat, Barley, Rapeseed, Beans, Peas, Vegetables, Fruit trees, Grapes…).

The Adaptive Visualisation Dashboard module is an advance enabler, which facilitates the dashboard DSS implementation. The technological choice is based on using Knowage as a visualisation and adaptive framework for building and developing the DSS user interfaces (business driven and not pilot specific).

This enabler facilitates the integration of AFARCLOUD compatible sensors and IoT devices into DEMETER ecosystem. The enabler subscribes to the MQTT broker where the observations are being published, converts the data into AIM model and stores the data in a triple store.

This component provides entry points to obtain weather forecast information. It exposes, using DEMETER AIM data format and a REST API, hourly weather forecast data (i.e. air temperature, relative humidity, wind speed, etc.) from trustworthy external sources (i.e. Openweather, Weatherbit, etc.) for a given geolocation. Note that time response will depend on their availability of the weather services.

The Estimate Beehive Component is a REST API designed to provide estimations of the pollination requirements for a given field. The application is a REST service designed for use by other clients registered in the Demeter Access Control System. To access and use the service will require a valid Bearer token obtained through the OAuth2 Client Credentials OAuth flow.

This component is designed to provide information related to the presence of sterile and non-sterile flies in automatic traps. It uses pictures taken under UV light that allow the insect counting. This counting process is based on automatically generated models from a set of labelled images (including examples of both types of flies). The component relies on the component “Pattern Extraction with Computer Vision”.

This component provides access to the necessary tools to train and use a crop yield model at the parcel level. The component works in close collaboration with VITO’s WatchItGrow database. The necessary yield data for training a model should be present in this database. For inference purposes, fields can be either extracted from the WatchItGrow database, or can be provided in AIM format to the service.

This component provides the analysis of any possible soil moisture anomalies within a plot based on satellite imagery.

Poultry Well-being Enabler evaluates overall stress levels in poultry by analysing environmental parameters (temperature, humidity, air flow, light intensity, and CO2) and video data patterns. The component presents a UI dashboard with air temperature, humidity, airflow, light intensity, CO2 level, animal species, detected stress levels, flock age, and safety instructions.

Component for Sensors Agronomic Information. This component, exposes agronomic information of a sensor deployed on a plot, registered in the pilot cloud infrastructure, modelled using DEMETER AIM. To do so, it calls an entry point in the pilot cloud infrastructure to retrieve the data that the user wants/needs to expose.

Transport Condition Enabler for poultry production covers the post-farm cycle and provides stakeholders with valuable insights into environmental conditions during transport. Input data, including latitude, longitude, CO2 levels, temperature, and humidity, is received in the DEMETER AIM format. After analysis, the classification output is provided to the dashboard.

The component provides the table with a list of drivers and list of machines with details about the driver behaviour, machine distance covered and vehicle average speed. The components receive data about latitude, longitude, speed, braking and fuels consumption as input in the DEMETER AIM format and the output is then provided to the dashboard (behaviour, distance covered and average speed).

This component presents the animal feeding quality based on food consumption and comparison with technology required consumption. The component provides the UI dashboard with food level in the silo with estimated consumption. The animal feeding consumption can be then assessed by farmers based on the level of the food in the silo. The components receive data as an input in the DEMETER AIM format.

The data management module block module consists of three main software sub-modules: ACS – Access control server DEH – DEMETER Enabler HUB BSE – Brokerage Service Environment Each of these modules exposes standard APIs that depending on the case, perform specific tasks in the data management process.

This enabler can take data from two sources and merge them by imputing missing data from one dataset to the other. It is also planned to be able to use limited predictions of future values in order to check for errors in data values. Enabler will be posted once all the functionality is debugged.

The Field Book and Fast includes both frontend and backend components. These components are designed to record various farm-related information, particularly phytosanitary treatments. The main goal is to assist farmers in maintaining digital field books that can adapt to different requirements. This resource specifically focuses on the backend package of this component.

This DEE component presents an optical trapezoidal model (OPTRAM) algorithm to infer the surface soil moisture with physical units using ML techniques and data fusion with remote sensing using satellite multispectral imagery and local sensing using ground soil moisture probes data.

This module is the implementation of the DSS dashboard. It is based on the Benchmarking enabler and on the Knowage framework. The DSS system provides indicators for Kotipelto farm and allows the comparison with neighbour milk farms. The monitoring of farms’ income and business activities is based on FADN European database.

Instance of Knowage (an Open-Source suite created by the Engineering Group that is focused in data mining/federation/visualization) has been used in order to display the Dashboard with the data coming from the component Pest Estimation with Sterile Fruit Fly. The component shows the number of captures of sterile and non-sterile fruit flies and allows the user to apply filters to such data.

This component estimates crop seeds amount to apply by different areas. Data analysis is done base on field coordinates and geometry and satellite images. The application will consider the dose rate and the number of seeds the user provides in the AIM input. The seed amount will be determined according to NDVI values (larger seeds for higher NDVI or lower amounts for lower NDVI).

This component will be used by the consumer in order to visualize the traceability phases on a user-friendly dashboard

Platform designed for the automation of various processes, whether they involve internal tasks performed by people, external tasks handled by software-based services, remote control of manufacturing machines, or a combination of these activities. DBT operates as SaaS model, where users can subscribe to different membership types to access a wide range of functionalities.

Web platform for water footprint estimation, water needs analysis and management of Cypriot vineyards. Specifically, it provides to farmers and producers in Cyprus information for water needs and footprint estimation of their vineyard products. Platform services are related to maximizing water efficiency and address water scarcity, as well as sustainable management of soil and water resources at field level.

Application to access Waisense Fields Data. It shows all the information about the plots and his linked devices.

Soil Interpolation using soil moisture sensors and satellite imagery combined with machine learning model is one of the innovative solutions of VineSignal Suite by DeepPlanet. It provides soil moisture content at all points and depths of the farm. The Web Portal provides user friendly way to access the Soil Moisture prediction videos to the farmers.

This enabler estimates ground salinity in areas without ground sensor data. The tool searches for salinity data within a specified timeframe, calculates average salinity using Inverse Distance Weighting (IDW) interpolation, and displays it with real earth coordinates in the application.

It is an application in which the user can input UAV/satellite images (RGB, Reflectance) and get an estimation of the soil salinity throughout the field using statistics models.

This platform for monitoring of working conditions of workers in food production company. Two different types of WiFi/Ethernet sensor nodes were deployed in a Food Production Company. 1) nodes to measure environmental temperature and humidity and module to monitor the air quality; 2) temperature sensor, to control de temperature in the oven’s exterior.

LivestockFeature dataset contains HealthPrediction data on milk production, fats, proteins, electrical conductivity, detected in the milk produced, the activity and rest of the cows, determined by the pedometer, the health status of the cows for the pathologies of lameness, mastitis and ketosis. MilkQualityPrediction: data relating to the analysis of milk samples and the degrees of milk quality.

This resource corresponds with the dataset of pictures of flies captured that is used in the pilot in order to create the data model that is used later on in order to calculate the number of flies in a picture.