Lookup NU author(s): Yu-Ching Lin,
Professor Jon Mills
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Small footprint, full waveform airborne loser scanning provides the opportunity to derive high-resolution geometric and physical information simultaneously from a single scanner system. This study evaluates the influence of various factors on the shape of the returned waveform and investigates the possibility of improving terrain classification by applying waveform-derived information. The factors discussed are surface roughness, slope angle, scan angle, amplitude, and footprint size. It is statistically demonstrated that roughness is the most significant factor affecting pulse width, and that, over relatively smooth surfaces, there is no significant variation in pulse width behavior resulting from different footprint sizes. Pulse width also exhibits a relatively stable behavior when amplitude, range distance, or scan angle vary substantially. The overall accuracy of classification achieved by applying pulse width information over all the different land-cover types examined in this study (including scrub, hillside, single trees, and forest areas) was greater than 85 percent, with >94 percent achieved for open vegetation areas. Physical surface information provided by small footprint waveform data is considered to be at the microscale, therefore it is recommended to combine such information with geometry (e.g., filtering algorithms) for the optimal identification of terrain points.
Author(s): Lin YC, Mills JP
Publication type: Article
Publication status: Published
Journal: Photogrammetric Engineering and Remote Sensing
Print publication date: 01/01/2010
ISSN (print): 0099-1112
Publisher: American Society for Photogrammetry and Remote Sensing