PROOF OF CONCEPT
Solve today's challenges to improve tomorrow's agriculture with us.
How do we proceed
1. Requirements analysis
What are your challenges and how can we help you to solve them ?
2. Target definition
We suggest a solution including sensors, methodology, acquisition protocols and analysis methods.
3. Test Campaign
We test the robustness of the methods with you in real conditions.
4. Validation cAMPAIGN
We report the preminilary results to you and we check the scientifical validity of our models.
We create the most optimal and easy-to-use solution so you can become autonomous with this solution and reach economies of scale.
Cercospora contamination was evaluated by a score between 1 and 9, representing the percentage of necrosis. It was done by technicians in the field, based on personal expertise and local knowledge. These scores will later be compared to the results of our analysis.
We defined the sensors (RGB and multispectral), the acquisition vehicles (UAV and Phenomobile) and the acquisition protocol (flight, speed and more…) in order to analyse each microplot sensitivity to Cercospora.
The data analysis involved a neural network computing multiple descriptors (size of nercrosis spots, Fcover, NDVI, Green Fraction…) and linking them to ADPC (Area under Disease Progression Curve), a useful quantitative indicator of the disease evolution.
Our test campaign showed that :
– Estimated ADPC allows better differenciation of different genotype classes than field scores.
– Estimated NDVI ADPC allows identification of sensible / less sensible genotypes
The amount of microplots studied had been scaled up from 100 to 1400 plots.
This phase allowed us to validate the methodology used to assess the evolution of the disease, and confirm the validity of the acquisition protocol.
During this phase we selected the solution that would trigger economies of scale and efficiencies for our client.
This solution was implemented on 4300 microplots.
Have a project you want us to be part of ?
Have a project you want us to be part of?
Hiphen was created in 2014 with the goal to make operational the results of years of scientific research about plant measurements and phenotyping in order to support the digital transformation of the agricultural sector… learn more.