8
THE DAVEY BULLETIN
|
September/October 2022
MISSION
MACHINE LEARNING
Davey Resource Group's (DRG) Environmental Consulting services (EC) is conducting
a street tree inventory for the city of San Diego. This is the first time DRG is using
machine learning throughout an entire inventory.
At the project's start, the city estimated they had around 220,000 trees. Jeff Harvey,
associate consultant and the street tree inventory's project manager, and Mike Bova,
area manager, ran a machine learning geographic information system (GIS) review of
San Diego's street trees. DRG utilized SiteRecon, an artificial intelligence (AI)-enabled
technology which uses Google Street View and aerial imagery data, to map the city.
They discovered San Diego had closer to 320,000 trees.
"We've really done a lot in teaching the machine how to identify trees within a certain
space from the curb," Bova said. "It's able to tell the difference between a light pole
and a palm tree, and uses a combination of heat, Google Street View, and aerial view,
to narrow down what a tree might be. SiteRecon is able to identify everything that's
more than 10 feet tall, so it's not going to count a shrub or a small fruit tree. This
technology also shows if there are no trees anywhere on a street. It saves time during
the boots on the ground inventory. It's been a great tool in determining where the
Left: From left, Parshuram Acharya and Jeff Harvey visually inspect a palm tree in a
San Diego neighborhood.
Right: Eric Meyer measures a tree's diameter at breast height (DBH).
