Course Dates
July 5, 2027 to July 19, 2027
Prerequisites
None
Course Description
This hands-on course introduces students to computer vision—using computer programs to identify,
interpret, and extract information from images and video—as a tool for studying marine and coastal
ecosystems. Students learn to design field studies, strategically deploy cameras and other imaging
equipment across diverse habitats, and analyze visual data using Python with OpenCV. The course
requires no prior programming experience and focuses on building an intuitive understanding of how
computer vision works. Lectures cover the basics of computer vision and its growing role in ecology,
behavior, and conservation research, including measuring predator-prey interactions, tracking collective
movement in fish schools and bird flocks, counting individuals to estimate population size, and detecting
behavioral shifts as early warning signs of ecological distress. Student-led projects focus on local biota,
including nearshore fish, seabird residents of Appledore Island, intertidal plants, marine invertebrates,
plankton, benthic cover, insects, and other coastal organisms.
Learning outcomes include:
- Apply the fundamental principles of computer vision to field biology.
- Design and execute field studies to address scientific questions and test hypotheses.
- Deploy and recover camera/video equipment in terrestrial and aquatic habitats.
- Use Python/OpenCV to perform basic image processing techniques, detect features and patterns
in images, and apply basic machine learning models to visual tasks using data collected from the
field. - Interpret and communicate research findings by connecting computer vision results to central
concepts in ecology, animal behavior, and conservation biology. - Critically evaluate the strengths and limitations of automated/computational methods used in
ecological field research, including the effects of data quality, collection conditions, and model
assumptions on results. - Determine the ways that automated / computational methods can be misused or misinterpreted,
including issues of bias, false detections, and overgeneralization of findings to broader ecological
conclusions.
Registration is CLOSED. Please check back in the fall for next summer.
Course Numbers
Cornell: BIOSM 2550 (3 Credits)
UNH: MEFB 585 (4 Credits)
Faculty
Andrew Hein