摘要: |
LiDAR (or airborne laser scanning) systems became a dominant player in high-precision spatial data acquisition to efficiently create DEM/DSM in the late 90's. With increasing point density, new systems are now able to support object extraction, such as extracting building and roads, from LiDAR data. The novel concept of this project was to use LiDAR data for traffic flow estimates. In a sense, extracting vehicles over transportation corridors represents the next step in complexity by adding the temporal component to the LiDAR data feature extraction process. The facts are that vehicles are moving at highway speeds and the scanning acquisition mode of the LiDAR certainly poses a serious challenge for the data extraction process. The OSU developed method and it implementation, the I-FLOW program, have demonstrated that LiDAR data contain valuable information to support vehicle extraction, including vehicle grouping and localizations. The classification performance showed strong evidence that the major vehicle categories can be efficiently separated. The I-FLOW program is ready for deployment. |