USE OF ‘MACHINE LEARNING’ TO DISTINGUISH BETWEEN TREE CANOPY AND TREES: “Trees are pretty hard to map. So, what’s the solution if we want to map tree canopies in places with complex geographies? How do we fill in the gaps between official street tree census and trees in parks and on private property?” – Tim Wallace, geographer for the New York Times

December 30, 2018 - Posted in Experience Around the World, Rainwater Capture: Planning, Tree Canopy Interception

Note to Reader:

In 2017 the San Francisco Planning Department finished a long-anticipated census of trees in San Francisco, and even launched an interactive map for identifying nearby leafy neighbors.

It’s a great tool, but the city only counted street trees on its Open Tree Map. The count didn’t include the likes of public parks, which meant that the majority of trees in San Francisco remained unaccounted for.

A more complete census would take longer—longer perhaps than anyone has time and resources to dedicate to the process of tallying trunks.

Machine learning might be able to tally trees that human labor simply can’t

Tim Wallace says that manually counting trees across a city is labor intensive. NASA uses satellite data to try to estimate vegetation canopy at specific locales. That’s a useful tool, but still not something that can easily provide an actual tree count. So Tim Wallace has drawn attention to a new approach by Descartes Labs, a satellite data company. Descartes uses artificial intelligence and machine learning as an arboreal abacus. The technology can read satellite images and other high-res scans and pick out which green bits are individual trees and which aren’t.

Mapping All of the Trees with Machine Learning

“Much fuss has been made over city trees in recent years. Urban trees reduce crime and help stormwater management (yay!). Cities and towns across the U.S. are losing 36 million trees a year (boo!),” wrote Tim Wallace.

“But, hold up—climate change is accelerating the growth of urban trees in metropolises worldwide (boo/yay?). Urban trees are under such scrutiny right now that the U.N. even had a World Forum on Urban Forests to discuss the planning, design and management of urban forests and green infrastructure.

“So Descartes Labs built a machine learning model to identify tree canopy using a combination of lidar, aerial imagery and satellite imagery.”

“The pattern of trees in any city reveals something about its urban planning history and legacy of greenspace. Don’t you want to know what your city’s treescape looks like?”

To Learn More:

To read the complete article, download a PDF copy of Using artificial intelligence to map every tree in SF

And then click on the link to read Mapping All of the Trees with Machine Learning