Davey Resource Group

POLLi's common milkweed identification models were tested in a peer-reviewed study published in Insects (MDPI) in 2026. POLLi collected the drone imagery and processed it through its machine learning models across diverse ROW types -- distribution lines, transmission corridors, gas lines, solar arrays, and transportation easements. The study included operational time and cost data alongside accuracy metrics, so land managers could evaluate UAV-based machine learning against four ground survey methods on the factors that actually determine whether a monitoring approach is practical at scale.

The published findings showed that POLLi's machine learning approach and complete site-level ground surveys most consistently quantified milkweed stem density across sites. Partial ground count methods showed greater variability depending on site and sampling approach. For land managers deciding which method fits their program, the study provides published, peer-reviewed evidence about where aerial machine learning performs well and where limitations apply. The CCAA began accepting remote sensing data in 2025, making this comparison directly relevant to ROW operators participating in or considering that program.

Davey Resource Group's scientific staff -- researchers in integrated vegetation management, biodiversity monitoring, and environmental consulting -- contributed the expertise that makes the study credible: the ground survey methodology, the field protocols, and the ecological knowledge needed to evaluate what the imagery-based findings actually mean on the ground. That kind of Subject Matter Expert grounding is what POLLi's evidence workflows are designed to preserve and carry forward, across monitoring cycles, sites, and the operational changes that any long-running ROW program will encounter.