Leafsnap: a mobile app to identify tree species

By June 7th, 2011 at 9:40 am | Comment

Leafsnap: By Columbia University, University of Maryland, and Smithsonian Institution

Leafsnap: By Columbia University, University of Maryland, and Smithsonian Institution

I love being in the outdoors amongst nature – but then who doesn’t? I also have a fascination for all things technological. Sadly, all too often these two passions are incompatible. For as us techie-lovers know, too many an hour can be spent cooped up inside staring at a computer screen.

The emergence of the smartphone now means that we can effectively carry powerful little computers around in our pockets. Programmers have sought to exploit this new technology and let citizen scientists get more involved.

Hot on the heels of MoGo and SoundAroundYou, Columbia University and the University of Maryland teamed up to create a new iPhone app called Leafsnap. Seeking to use smartphone technology to engage people in their environment, it promises to answer to that question, “I wonder what type of tree that is?,” when you don’t have anyone to ask.

Utilizing visual recognition technology and an Internet-enabled smartphone, the Leafsnap app identifies plant species with the phone’s built-in camera. I was excited by this prospect, and so tearing myself away from the laptop (iPhone in hand), I set out into the great outdoors to put Leafsnap through its paces. Here’s what I found:

Leafsnap performed well on both an iPhone and iPad; it is easy to use and boasts a wealth of great features. After “snapping” and uploading an image of a tree’s leaf, you are presented with a list of likely candidates. Bark, flower and leaf images and accompanying facts then let you work out if Leafsnap has found your tree. Your findings are saved and placed on a world map, letting you see what other people have also spotted in your area. I never knew the Yoshino Cherry tree grew in our part of the world!

I also found that leaves need to be taken off the tree and then photographed on a plain white background. So unless you’re going to carry a white piece of card around with you, it works best if you take your leaf samples home. Even with a good mobile signal, recognition time for each specimen takes at least a minute, which might put off the impatient amongst us.

Of the five leaves I tested (Walnut, Crab Apple, Oak, Horse Chestnut and Lime tree), Leafsnap got three of them right. Given the subtle differences between many leaves, this in itself is pretty impressive. As an aid to tree identification and a repository of useful facts and images, Leafsnap is incredible. A foolproof solution to tree identification, this is not. Nevertheless, if Leafsnap gives you an excuse to get outside exploring nature, then this can only be a good thing.

Sign up for the Science for Citizens newsletter!

This guest post was contributed by Dr. Stuart Farrimond, a science teacher at Wiltshire College in the United Kingdom. Check out all of Dr. Stu’s Reviews!