Individual Tree Location and Canopy Delineation Based on Quickbird Imagery

Remote sensing data with high spatial resolution can potentially be used to estimate the biomass of individual tree crowns. As a first step, the location of each tree must be identified; this step is particularly important in dense forests. Total station (Pentax R-400) and a handheld differential Global Positioning System (GPS) were used in tandem to determine accurate locations of single trees. The locations were then combined with high spatial resolution remote sensing data to extract additional forest vegetation information. Tree location data were interpreted visually and were also the inputs for a ray-based automated method of crown delineation. Compared to field-collected crown data, the mean accuracy of the visually interpreted data was 65% in plot B3 (RMSE = 0.60) and 79% in plot B15; plot B15 did not have a dense canopy and had a Smaller Error Statistic (RMSE = 0.32). The ray model was less accurate. Crown size estimated from both the visually interpreted data and the ray-based model was usually smaller than that estimated from field data. One conclusion of this study is that crowns with a narrower density distribution are correlated with a higher error rate. The crown delineation method proposed in this study was shown to be feasible.

並列關鍵字

Canopy ； GPS ； individual tree ； remote sensing ； total station

國際替代計量

全文下載

主題瀏覽