Our team of experts supports you from proof of concept to the final product, for the development, testing, integration and use of powerful models for the prediction and simulation of plant growth in an agronomic context.
We combine the theoretical and historical approach of scientific crop models with artificial intelligence, machine and deep learning. The former provides scientific knowledge, the latter learns from available data to improve prediction by contextualising it.
To go further, we take the best of two worlds.
Our experts have worked on complex agronomic models developed in a research context: APSIM, WOFOST, CERES and STICS. They are used to developing forecasting modules for different plant species for concrete agronomic applications. Moreover, they are trained in the latest machine and deep learning techniques, both for the integration of sensor data for calibration and model validation, or for the coupled use with crop models to refine and sharpen predictions.
Modelling to anticipate the risks that apply to crops and understand the combined effects of climate, soil and technical itinerary in order to assess and evaluate potential yield losses and damages and make the right decisions at the right time.
- Assess the risk of drought for crops at local or regional scale.
- Estimate the risk of disease outbreaks.
- Predicting crop yield prior to harvest.
Modelling to evaluate what cannot be measured is a fundamental approach for varietal selection and agronomic research. We estimate variables that are not accessible or not directly measurable, but are fundamental for characterising environments and varieties in a changing climate.
- Estimate the water use of a variety, by calibrating the model from soil measurements (e.g. tensiometer) and leaf area measurements by camera or UAVs during the growing period.
- Characterise a network of fields as close as possible to the stresses experienced by the plant: envirotyping.
Upstream of real experimentation, modelling can help simulate and test new technical itineraries, new practices, new varieties, or new environments. This allows hypothesis testing and clearing the way for cost-effective experimentation.
- Testing virtual varieties to target new traits of interest that maximise production: ideotyping.
How does our service work ?
You have a problem, we have the methodology to adress it. Our team is involved in all the key stages to achieve your objectives
Is this possible? We study the existing situation and set up a POC (proof-of-concept) to study and evaluate the feasibility of a project.
What data do I need? We work to define and process the data to be measured in order to calibrate and validate a reliable and relevant solution. If you already have the data available, we treat it, else, we help you define and measure what you need.
Which solution, which model, which methodology to use? We choose the best solution, adapted to your project, if it exists we help yousetting it up, if it does not exist we develop it for you, with you.
How to integrate this solution in my existing environment? We help you integrate the solution, existing or developed with you, into your hardware and software environment. If necessary, we can provide you with our expertise to set up the right environment.
Your problem is unique, contact us so that we can answer your questions and make a proposal!