A Remote Sensing Based Urban Tree Inventory for the Mississippi State University Campus

A unique urban tree species identification project for the Mississippi State University campus was undertaken in the Spring of 2002. A collaborative effort between the MSU Forestry Department, Landscape Architecture Department and the U.S. Forest Service was initiated in an attempt to automate tree species identification from high-resolution multi-spectral imagery. Traditional image processing approaches to tree identification failed to accurately classify trees into unique classes due to differential crown illumination, crown shadowing, and spectral confusion with other vegetation types. Feature Analyst enabled extraction of tree crowns from other landscape features. After tree crowns were delineated from other vegetative cover types, a traditional image processing approach yielded excellent results for delineating Pine, Hardwood, Cypress, and Other tree species.