Abstract
In precision forestry, tree species identification is one of the critical variables of forest inventory. Lidar, specifically full-waveform Lidar, holds high promise in the ability to identify dominant hardwood tree species in forests. Raw waveform Lidar data contain more information than can be represented by a limited series of fitted peaks. Here we attempt to use this information with a simple transformation of the raw waveform data into the frequency domain using a fast Fourier transform. Some relationships are found among the influences of component frequencies across a given species. These relationships are exploited using a classification tree approach to separate three hardwood tree species native to the Pacific Northwest of the United States.
We are able to correctly classify 75% of the trees ( 0.615) in our training data set. Each tree's species was predicted using a classification tree built from all the other training trees. Two of the species grow in proximity and grow to a similar form, making differentiation difficult. Across all the classification trees built during the analysis, a small group of frequencies is predominantly used as predictors to separate the species.
Acknowledgements
The authors thank Terrapoint USA Inc. for providing the data set used in this study, as well as the University of Washington Precision Forestry Cooperative, the University of Washington Remote Sensing and Geospatial Analysis Laboratory for their support. They also thank Bob McGaughey and Steve Reutebuch who provided their experience and their assistance in working with the raw and processed data at the USDA Forest Service Pacific Northwest Research Station.