A project to automate the counting of glacier ice worms from photos of snowfields.
A friend and I are trying to improve the field method for counting ice worms on glacial snowfields. Most (or all) field studies so far have involved researchers marking out square-meter plots and counting all ice worms in the plot by hand. But ice worms tend to appear for a short period at sunrise, and then retreat back into the snowpack during the day.
We plan to photograph the snowfield in a short period of time, and then process the images for an automated count of all the worms on a snowfield. We will focus on the snowfields of Baranof Island in southeast Alaska. As far as we can tell so far, no one has studied the ice worms on Baranof Island.
So far, the program finds all the pixels that appear to be ice worms. The next step is to examine these pixels and come up with a count, and then compare that to a hand-count for a small number of reference images.
The project is named after a children's book that a friend gave me a few years back, after we first encountered ice worms in the mountains of Baranof Island.
If you'd like to get in touch about this project:
@ehmatthes