In this seventh episode of a series of CAPTE videos, Joss Gillet from Hiphen and Samuel Thomas from Arvalis talk about Cloverfield, our online data processing engine built to automate the processing of plant phenotyping datasets.

A secured and cloud-based architecture

Samuel explains that our data platform is built on a secure cloud-based architecture running on Amazon Web Services and docker technology to efficiently assemble and trigger the various algorithms required to process specific agronomic traits for our clients.

This online tool can process data acquired from any sensors, such as UAV (drone), Phenomobile, satellite, IOT and handheld devices such as Literal. This type of data engine is critical for any agro-actors such as plant breeders, farmers and cooperatives that have large fields or microplot trials spread across vast regions across the world. For them, it is very costly to compute all the data acquired on their fields and trials, and results tend to take a long time to be delivered. In addition, it often limits the amount of data clients think they can acquire, and they tend to prioritize certain fields or trial experiments to the detriment of others because processing too much data was up to now almost impossible, too expensive or would have required to in-house technical skills that would have made the task daunting.

 

How it works?

First of all, Cloverfield allow to collect crops images after the data acquisition. After that, the automated data processing starts to run thanks to our algorithms and deep learning techniques. Then once the processing is finished, the next step is data visualisation as you can browse the agronomic traits computed very intuitively on the field map.

A new dimension for plant phenotyping

Cloverfield removes all these barriers and brings a new dimension to the plant measurements and phenotyping ecosystem, allowing users to concentrate on acquiring data, knowing that with our solution we can process large volume of data in a matter of days. The tool is flexible enough to accomodate for the specific agronomic traits selected by the client, but it can also deliver direct or intermediary outputs such as raw images, co-registered images, orthomosaics, microplot extractions, etc.

We use Cloverfield to deliver global contracts for our clients, ranging from plant breeders to agro-industrial actors with an international footprint. For instance, we receive UAV data all year long from countries in the Americas to Europe and Asia that Cloverfield can then process in a very timely manner.

Do not hesitate to get in touch with us should you wish to learn more about Cloverfield and how to gain access to it. We look forward to hearing from you.