Founder: Romain Fau
Measurement, Reporting and Verification (MRV) is a required part of any climate project. However, conventional MRV approaches for nature-based solutions require field visits by forestry experts or specialized consulting firms. These methodologies are expensive and specific to each project, take many months or even years to compile data, meaning economies of scale cannot be achieved. This often makes them prohibitive for small and medium sized projects. Results are also heterogeneous and use distributed sources, creating a lack of transparency and as a consequence a lack of trust on carbon markets.
Kanop is a SaaS for Measurement, Reporting, and Verification (MRV) of nature-based projects. With a primary focus on forests, we apply proprietary machine learning models to satellite images of forests (optical and SAR) to give indicators up the singular tree level, such as species, height, and carbon sequestered, to name a few. Our self-service web app can deliver results quickly, accurately, and transparently, allowing users to aggregate and view data as they choose. This digitized and automated solution allows project developers and other stakeholders in the carbon market to scale their MRV activities without cost overruns, bringing transparency and trust to all stakeholders without excluding the essential small and medium-sized players.
Kanop works on a subscription model. Kanop’s services are divided into six packages, with pricing per hectare decreasing as the measured area increases. Within each package, project developers can mix and match measurements at either 10m or 25m resolution based on their specific requirements.
ML technologies are central to Kanop’s products. This involves developing different processes:
– Image acquisition: identification, downloading, and pre-processing of satellite images of the area of interest (optical and SAR). This step is crucial because the required image specifications vary according to regions and seasons in the area of interest. Once acquired, the images are pre-processed upstream of the models with normalization, filtering, and possible registration operations between images.
– Models: to derive the most relevant and explainable forest indicators, Kanop combines different machine learning / deep learning models
– Post-processing: Kanop processes the model-produced data to correct residual errors. This involves cross corrections between different models as well as common sense rules that ensure the consistency of the delivered data.
After 18 months of R&D and pilots, our adoption is now scaling with 15+ paying customers.
Romain Fau has 10+ years of experience in the technology industry. As the General Manager for Western Europe at BlaBlaCar, Romain has been instrumental in managing BlaBlaCar high growth from 2012 to 2017. More recently, Romain led product and monetization at PassFort (now Moody’s Analytics) and Cleo AI. Romain co-founded Kanop in 2021 with Louis, our CTO, with the idea of bringing transparency and trust to the carbon and biodiversity markets. Louis de Vitry, Co-founder and CTO, holds degrees from EPFL, CentraleSupelec, and KTH. He has managed various AI projects, including multi-modal image analysis for cancer treatment improvement.
As global focus on climate action intensifies, several regulations, including the forthcoming European Union Due Diligence Regulation (EUDR) and Corporate Sustainability Reporting Directive (CSRD), are poised to shape the landscape of environment.