Light Detection and Ranging (LiDAR) provides remotely sensed information describing the vertical structure of forests. Using LiDAR-derived metrics such as mean canopy height as input to predictive models, additional forest metrics such as timber volume and biomass can be estimated rapidly and frequently for individual holdings. In this study, Optech's ALTM 1225 airborne LiDAR system was flown over a tolerant hardwood forest in the Turkey Lakes Watershed near Sault St. Marie, Ontario, Canada in August 2000. The focus of this study is a one-hectare area where mensuration data and coordinate locations for 675 trees were collected and the crown architecture of 171 trees mapped. The goal of this study is to relate laser tree height estimates with field tree height measurements for individual trees at a local scale. A "manual" technique where the user interactively analyzes the LiDAR data and a delineated forest canopy produced from a Laplacian of Gaussian (LoG) filter applied to a canopy height model was adopted because of the failure of previous experimental automated methodologies to produce expected results. Based on this technique, a linear regression produced an R~2 value of 0.92 that is significant at the 5% level. Furthermore, problems associated with the automatic classification of LiDAR data for forestry applications are discussed. The results of this study demonstrates the utility of small-footprint, time-of-flight LiDAR for estimating forestry-related metrics in deciduous forest ecosystems and shows that a tendency to underestimate tree heights does not exist as is the case in coniferous forest studies using LiDAR.
展开▼
