OUR APPROACH
Technology proportionate to the need
We put our knowledge to work and propose technical solutions to environmental and civic problems. Our technical perspective makes methods more robust, more effective and less costly, and opens up new use cases.
AI & ENVIRONMENT
What can artificial intelligence do for the environment?
Although the explosion of AI and its massive use pose a potential problem for the environment, we believe that a subset of this technology can be put to good use.
Better diagnosis
- Automatic counting and processing of litter and air or water pollution (see Cobra Propreté and Plastic Origins)
- Study of macro- and microscopic biodiversity (see the Karibou Project)
- Monitoring and anticipating climate change, deforestation and land take
Optimising what already exists
- Optimising energy costs, particularly for local authorities and industrial buildings
- Identifying waste of resources such as water and raw materials
FRUGALITY
Why frugal technology?
Adding technology, which is sometimes complex, always comes at a cost: it drives the use of ever more sophisticated hardware, consumes additional electricity and can make a system harder to understand. We therefore use frugal technologies wherever possible, which we define according to three criteria:
01
Energy-efficient
Favouring models with low computing and memory requirements.
02
Easy to understand and maintain
Open source where possible, which makes them more accessible to all.
03
Hardware-light
Building on existing capacity for deployment, rather than pushing for the purchase of new devices or servers.
TECHNOLOGY
What technology do we use?
We master the entire information processing chain: storage systems, databases, application backends, mobile apps and frontends for data visualisation, especially mapping. We work mainly in Python, relying on open-source technologies.
When it comes to AI, we use classic data analysis methods as well as modern techniques such as deep neural networks. We have extensive experience in the following areas:
- Computer vision : object detection and segmentation in images, object tracking in video
- Cloud AI : deployment and infrastructure that optimise computing costs
- Embedded AI on mobile (Android and iOS)
Do you have an environmental project?
Need to turn environmental data into a simple, robust and frugal digital tool?