Photo: Leafsnap
Tree enthusiasts are always on the lookout for new tools for identifying trees, and this time, it’s free. Leafsnap is a new iPhone app (yours is coming soon, Android users) that allows you to snap a photo of a leaf and have it instantly evaluated via image-recognition technology to find the closest match among its database of select trees from the Northeast.
On a recent day in Arlington, VA, I put Leafsnap to the test. Trying to snap a photo of the leaf while it was still on the tree proved to be a logistical challenge as the leaves need to be photographed on a white background. So, leaves from a tulip poplar (Liriodendron tulipifera), Japanese red cedar (Cryptomeria japonica), kousa dogwood (Cornus kousa), and Southern magnolia (Magnolia grandiflora) were first plucked and then snapped.
The results? It’s not 100% yet but the Leafsnap app shows a lot of promise.
At first, Leafsnap thought the cedar was a maple but better lighting and photography helped it to immediately recognize it for what it was. The tulip poplar also came back as the first match. The kousa dogwood was a little harder as there are a number of other trees that have similarly shaped leaves but it still came up in the top three results.
The Southern magnolia was the trickiest. The first snap came back as entry number 10 on the list but turning the leaf over and capturing the back side of the leaf gave much better results – number 2 on the search results list. The bottom line – and somewhat to a thinking-person’s relief – you still need to use your head to make the final determination.
For the casual tree browser in you, Leafsnap allows you to peruse photos of the trees currently in the system – more expected soon. For each entry there are photos of the leaves, flowers, fruit, petiole, seeds, and bark, and a brief description. Sitting on the bus on the commute into work — often at leaf level of those small city trees, endless minutes can be spent scanning Leafsnap’s list for possible matches.
And then once you have a handle on all of the database’s trees, try your hand at the built-in games to see how leaf literate you are. Leaves are one thing, but identifying fruit is tricky. These photos are close-ups of the fruit – a vantage point you don’t often get from the ground. Suffice it to say that the author’s Leafsnap game scores are not going to be reported.
Developed by researchers from Columbia University, the University of Maryland, and the Smithsonian Institution, they have big plans for taking Leafsnap national. At the moment, they are inviting citizen scientists to submit leaf images, which then get geotagged and added to a collection on the leafsnap.com web site. The web site also has a whimsical gallery of the people behind the app – each one holding a leaf.
As a TreeSteward answering questions at a farmers market tree table, you can imagine how handy it would be to have Leafsnap in your pocket. Just about every tree project begins with the question, “What kind of tree is this?” Whether planting a tree, pruning a tree, or watering a tree, first identifying the tree is the key to figuring out how best to handle the situation – and Leafsnap promises to be a great aid.
