Interpretation of upper-storey canopy area in subtropical broadleaved forests in Okinawa Island using lase scanning data

Conventional forest inventory practice took huge of effort, and is time- and cost- consuming. With the aid of remote sensing technology by light detection and ranging (LiDAR), those unbearable factors could be minimized. LiDAR is able to capture forest characteristic information and is well known for estimating forest structure accurately in many studies. Forest monitoring related to forest resource inventory (FRI) becomes more effective by utilizing LiDAR data and it is tremendously useful, especially to distinguish information on density, growth and distribution of trees in a selected area. In this study, LiDAR data was utilized aimed to delineate crown cover and estimate upper-storey canopy area in Yambaru Forest using object-based segmentation and classification techniques. Agreement between field survey and LiDAR data analysis showed that only 33.7% of upper- storey canopy area was successfully delineated. The low accuracy level of canopy detection in Yambaru Forest area was expected mainly due to tree structure, density and topographic condition.