Computer Vision • Environmental Science • Research Infrastructure
Green Crossing AI is an interdisciplinary research project at Southern Oregon University that develops new computer vision methods for wildlife monitoring and conservation research. Bringing together computing and environmental science, we investigate how data science and modern computing can help researchers better understand wildlife movement through long-term ecological observations.
About
Wildlife crossings are becoming an important tool for reconnecting fragmented habitats and reducing wildlife–vehicle collisions. Understanding how animals use these landscapes requires long-term monitoring before, during, and after changes to the environment. Green Crossing AI develops computing tools that help researchers collect, organize, and analyze ecological observations, supporting conservation science and a deeper understanding of wildlife movement over time.
Project News
July 2026
Federal funding has been released for a wildlife overcrossing spanning Interstate 5 in the Cascade-Siskiyou National Monument south of Ashland. The planned structure will be Oregon’s first wildlife overcrossing and the first along Interstate 5 between Mexico and Canada.
Green Crossing AI supports the long-term research needed to better understand wildlife movement before, during, and after construction.
Research
Green Crossing AI brings together computer science and environmental science to develop methods that support long-term ecological research.
Understanding wildlife movement requires observations collected over months and years. We develop methods that help researchers organize, analyze, and interpret long-term ecological data to better understand changing landscapes and wildlife behavior.
Computer vision and data science provide new ways to process and analyze large collections of ecological observations. Our research investigates methods that help scientists extract meaningful information from camera-trap data while supporting reproducible research.
Green Crossing AI provides undergraduate students with opportunities to contribute to interdisciplinary research through software development, data analysis, field observations, and collaboration with environmental scientists.
Open Science
We share publications, software, and educational materials to support reproducible research and collaboration.
PEARC ’25 Conference Paper
This paper presents a practical camera-trap processing pipeline designed to help environmental science research groups manage large image collections and incorporate computer vision into their research workflows.
NAIRR ’26 Conference Poster
Presented as a poster at the NAIRR (National AI Research Resource) annual conference in March 2026, this work was part of the NAIRR Classroom Pilot award and highlights a reproducible camera-trap workflow for ecological research and wildlife monitoring. It also explains a module used in CS 317 Introduction to Data Science.
Open Notebooks
A collection of open Jupyter notebooks for camera-trap processing, MegaDetector, computer vision, and data science workflows.
From the Field
Camera traps provide a window into wildlife activity across Southern Oregon. These observations contribute to a growing long-term record of wildlife movement and ecological change.
Image credit: Wildlife photographs are from camera traps deployed by Dr. Karen H. Mager and collaborators as part of ongoing wildlife monitoring in Southern Oregon.
Team
Green Crossing AI is an interdisciplinary collaboration involving researchers and students from Southern Oregon University.
Assistant Professor of Computer Science
Southern Oregon University
Associate Professor of Environmental Science
Southern Oregon University
Researcher and project collaborator
Current student contributor: Katherine Nunn.
Past student contributors include Erik Harden ’24 and Harley Chappel ’24.
Contact
We welcome conversations with researchers, students, conservation organizations, and community partners interested in wildlife monitoring and environmental research.