Spring NEARC 2019 has ended
Welcome to the interactive web schedule for the 2019 Spring NEARC Conference! For tips on how to navigate this site, visit the "Helpful Info" section. To return to the NEARC website, go to: www.northeastarc.org/spring-nearc.html.

The schedule is subject to change (as of May 13, 2019). Please check back for updates.
Back To Schedule
Tuesday, May 14 • 8:00am - 5:00pm
POSTER: Investigations on the use of Unmanned Aerial Systems for Collecting Thematic Map Accuracy Assessment Reference data in Complex Natural Environments.

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

AUTHORS: Benjamin T. Fraser* and Russell G. Congalton

ABSTRACT: With the expansion of modern technologies, remote sensing and geographic information systems (GIS) are able to capture and analyze data at ever progressing scales. These advances help todays? users create vital land cover and land use (thematic) maps using novel classification methods to represent increasingly complex environments. For these resulting thematic maps to serve as proper decision support tools for research and management, their accuracy must first be evaluated. The methods for assessing thematic accuracy have developed considerably over the years, now advising site-specific multivariate analysis using an error matrix. Despite improved methods of analysis, immense costs and time restraints on the collection of samples used as a standard of comparison for reality, reference data, often limit accuracy assessments. Many projects have high-resolution remote sensing, ground-based sampling, or maps of known accuracy as reference data, with ground-based reference data being the most sound yet costly. The relatively recent proliferation of Unmanned Aerial Systems (i.e., UAS, UAV, or Drone), with high spatial, spectral, and temporal resolutions may help to overcome this challenge. Our research at the University of New Hampshire analyzed first the ability to collect data of sufficient comprehension, in New England Forests. Next, we conducted a pilot study over 377.57 ha of woodlands that achieved 71.43% and 85.71% agreement to ground-based samples under pixel-based and object-based classifications respectively despite noted sources of uncertainty. Future applications and research objectives are briefly discussed to further encourage the use of emerging technologies as tools for providing information at management needed scales.

Tuesday May 14, 2019 8:00am - 5:00pm EDT
STUDENT CENTER: 1st Floor Lobby (Appian Way Entrance)

Attendees (2)