Geospatial AI Researcher & MS Student · Nelson Institute for Environmental Studies · University of Wisconsin–Madison
I'm a researcher and MS student at the Nelson Institute for Environmental Studies at UW–Madison, with a background in computer science and environmental studies. My work sits at the intersection of computer vision, remote sensing, and land system science, researching things such as AI models best suited to detect cattle linked to deforestation in the Brazillian Amazon, and the potential benefits to rewilding lawns in the United States.
My passion lies in working toward land system solutions that balance productivity, conservation, and livelihoods by making these tradeoffs more measurable through technology. Land use decisions are complex and involve different priorities, often favoring productivity. I'm drawn to finding that balance, and I see computer vision and Earth observation AI as the most direct toolkit for doing it at scale.
Built an inference pipeline to benchmark multiple super-resolution models on satellite imagery, comparing tradeoffs in predictions and runtime across architectures.
An internal tool for fast labeling of imagery, built to speed up the lab's bounding box annotation workflow.
A set of in-lab geospatial utilities, including a coverage-finder tool that checks which imagery providers have usable satellite coverage over a given area and time window, plus supporting spatial tooling for the lab's analysis pipelines.
Computer vision and geospatial tooling for satellite imagery analysis.
November 6, 2025
Early findings of cattle detected in embargoed areas of Pará, Brazil
Reach out via the form below, or feel free to email me directly at palominolina@wisc.edu