Vegetation Classification Using LiDAR Data
Event Type
DescriptionGlobal climate change and sustaining urban growth, have increased the demand for determining the role of urban vegetation in urban residents' lives, as well as in the ecosystem services. LiDAR is an optical remote-sensing technology for high-resolution and highly accurate topographic data acquisition, which uses laser light to densely sample the surface of the earth.

With the emergence of Airborne LiDAR, scientists were able to facilitate the development of tree species identification in forest ecosystems. Successful tree species classification with remote sensing data and using it for reconstructing the structure of objects such as trees after major natural disasters like hurricane or wildfire is valuable to forest inventory and ecosystem management.

To fill this gap, we propose to develop an ML model that applies the level set and Random forest algorithm combined to classify urban vegetation in the Apalachicola, Fl. Using the Digital Elevation Model (DEM), and Digital Surface Model (DSM) raster, point cloud with a 0.5 m resolution, we smooth the noises, extract the sinks by finding the depressions on DEM’s and perform numerical computations which involve curves and surfaces on a fixed cartesian grid and find a hyperplane in a 3-D space which distinctly classifies the data points and different vegetations.

The result of this study is essential for determining differences in urban ecosystem services related to preventing wildfire and the damages that the urban residents receive during the hurricane, Therefore, in-depth analyses of urban vegetation is very useful for global climate change detection.