USE OF LIDAR DATA TO AID IN DISCRIMINATING AND MAPPING PLANT COMMUNITIES IN TIDAL MARSHES OF THE LOWER CONNECTICUT RIVER: PRELIMINARY RESULTS

摘要

This paper summarizes the status and preliminary results of lidar research currently underway at the Center for Land Use Education and Research at University of Connecticut. Researchers are investigating uses of an October 8, 2004 lidar dataset that covers approximately a 300km2 coastal area of Connecticut from New Haven Harbor to the mouth of the Connecticut River. Lidar data are being used to aid in the classification of tidal wetland plant communities, which is being done through classification of high-resolution Quickbird satellite and airborne imagery using eCognition image segmentation software. Plant species of primary interest include Phragmites australis, Typha spp. and Spartina patens. These three plant species have distinct height differences that were expressed at the time of the lidar data collection. Boundaries of relatively pure stands of each plant species manually were digitized from 0.5m resolution ADS40 airborne CIR imagery and were used to analyze specific plant community subsets of the lidar data to determine lidar height characteristics. Mean lidar-based heights for the species of interest were: P. australis = 3.01m; Typha spp. = 1.37m; S. patens = 0.64m. The data also are being examined to learn how lidar interacts with these tidal marsh plant communities, to characterize transition zones (edges) between plant communities and to characterize mixed vegetation communities. Results of the analysis will be used to improve image segmentation and classification of multi-temporal Quickbird imagery, especially in locations where spectral properties of Quickbird's four bands may be insufficient to adequately discriminate among plant communities.