Feature Extraction
Using LiDAR data, digital orthophotos, and existing vector layers, we can delineate a wide variety of terrain feature classes that may be desired for a project. Commonly extracted features include building footprints and heights, bridges, levees, paved/unpaved roads, stream channels, ditches, tree canopy, and wetlands. Our overall approach to feature extraction involves the direct use of the raw LiDAR points as opposed to using intermediary surfaces (rasters) as primary sources. Features are classified through a series of automated and manual processing steps, with the ultimate product being a GIS-compatible 2- or 3-D vector layer with relevant feature attributes.
At right, point cloud showing classified building points in blue and vegetation points in green, with corresponding vectorized building data on left.
Paleo channels can be manually (visually) identified and delineated from the hillshaded bare earth LiDAR data.