Polli Announces Release of Habitat Index V2

Polli is excited to announce the release of an updated Habitat Index model. The latest update boosts overall accuracy by 13% and improves performance by up to 27% across various vegetation types.

This new version was built using more than 2,500 carefully reviewed images, providing the model with a better understanding of how plants appear across real, diverse landscapes. Unlike models designed for uniform farmland, POLLi’s Habitat Index must recognize the complexity of nature where trees, grasses, and flowers overlap, change with the seasons, and differ by region.

By expanding the training data, it now does a better job of distinguishing between similar vegetation types and identifying subtle transitions across habitat zones.

Key Improvements

• +13% overall accuracy and up to +27% gains in new vegetation classes

• Stronger detection of floral resources and native vegetation

• More consistent habitat mapping across diverse landscapes

• Improved confidence in results for reporting and planning

We continually improve the Habitat Index to support new ecosystems, monitoring contexts, and geographies, ensuring accurate results for users.  As new imagery and field data become available, POLLi continues to learn, helping you see habitats more clearly and manage them with greater confidence.