Proving accuracy with a new benchmark

Kanop uses advanced AI, remote sensing, and agronomic expertise to map forest metrics globally with high accuracy. Our platform combines deep learning models trained on various remote sensing data with extensive field data and expert knowledge. Independent assessments show our biomass and tree height estimates are comparable to NASA's GEDI data and outperform recent models.

Kanop is pioneering a new frontier in ecosystem monitoring by combining cutting-edge AI, remote sensing data, and agronomic expertise. Our innovative SaaS platform provides highly accurate, high-resolution mapping of key forest metrics like biomass and tree height anywhere in the world.

At the core of Kanop's approach are advanced deep learning models trained on a massive repository of multi-sensor remote sensing data - including optical imagery, SAR, and LiDAR from sources like Sentinel, GEDI, and others. But what truly sets us apart is the integration of extensive field data and decades of expert agronomic knowledge from our team of scientists.

We recently completed a rigorous independent accuracy assessment comparing our model’s estimates against global biomass and height datasets across 300 diverse 1km2 forest areas. The results validate Kanop's industry-leading performance:

Biomass Accuracy

• No statistically significant difference from NASA's GEDI published biomass data globally or at the biome level

• Mean Absolute Error of just 15.9 tons dry matter/hectare

• Outperforms recently published models like Wang et al. 2021 (RMSE 29.8 vs 22.1 for Kanop)

Tree Height Accuracy

• No significant difference from GEDI raw height metrics

• Mean Absolute Error of only 2.9m, near LiDAR accuracy

• Approaches state-of-the-art models like Staben et al. 2018 (RMSE 2.8m) and Li et al. 2020 (2.64m) which are not global models

What allows Kanop to achieve this unprecedented global accuracy? A few key advantages:

1. State-of-the-art AI models effectively leveraging multi-source remote sensing data including optical, SAR, and LiDAR
2. Unique training/validation datasets incorporating extensive ground data
3. Decades of agronomic expertise integrated into the data pipeline
4. Ability to create globally consistent, high-resolution 10m outputs

By pushing the boundaries of AI for environmental analysis, Kanop empowers organizations to precisely measure and maximize nature's climate benefits across any geography. Our team will continue advancing these capabilities to provide increasingly intelligent ecosystem monitoring solutions.