Computer vision · field tools
LitterCam
A camera-assisted litter recording system that connects real cleanup walks with structured environmental data.

The challenge
Recording litter while collecting it is slow, inconsistent work. The useful data should be easier to capture without getting in the way of a cleanup.
The work
Camera hardware, an iPhone workflow, and a computer-vision training and review process. Real field walks inform the models, with human review used to check what was detected.
What it demonstrates
Connecting physical work to digital records, testing automation in real conditions, and keeping human review where it matters.