Greenwich Institute for Science and Technology
   Home >> LIDAR Technologies >> 3D Terrain Model and Intensity Imagery
3D terrain model and intensity imagery:

      Virtual Surfaces Method for DTM generation and foliage filtering from LIDAR data is developed by scientists at the Greenwich Institute for Science and Technology. The Virtual Surfaces Method is a local approach that uses analytical surfaces for local fitting of earth surface. Process of foliage filtering is integrated together with bare earth surface interpolation in filtered pixels. The Virtual Surfaces Method have benefits of popular methods for DTM generation like Vosselman-Sithole and Kraus-Pfeifer techniques, but avoid most problems of these approaches (see details in GDATP/GIST Technical Paper "Terrain modeling: mathematical methods and numerical algorithms to describe terrain and surface features", 2004). High accuracy and other benefits of the Virtual Surfaces Method were tested during extensive industrial application for processing LIDAR data and DTM generation in area of ~25,000 (see Fig.1 and below).

First Return

Fig. 1. (Top). The Digital Surface Model with 2-m pixel for MD tile is 1800x1200 m (MD DNR data). (Bottom). The bare earth surface (DTM) generated by using Virtual Surfaces Method with accuracy 10-15 cm. All surface objects like residential buildings or trees were successfully removed.

Recent LIDAR county-size projects completed by GIST team:

Galveston, TX (~600 sq.mi. Visual Intelligence data, 3-m resolution)
Duplin, NC (~1000 sq.mi. Airborne-1 data. 2-m resolution)
Puerto-Rico coast (~1,500 sq.mi, NASA data, 2-4-m resolution)
Beaufort, SC (~700 sq.mi, Laser Mapping Specialists data, 1.2-m resolution, RMSE=6 cm)
Boston, MA (~120 sq.mi, 3D urban modeling, 3DI data, 1-m resolution)
Prince George's, MD; Mecklenburg, NC and data for TX and SC (FEMA project for 3D modeling from LIDAR data from 4 vendors; ~500 sq.mi)
Worcester, MD (~700 sq.mi, 3DI data, 2-m resolution)
Eastern MD (> 6 counties, see map below; ~3,000 sq.mi, Airborne-1 data, 2-m resolution, RMSE=14 cm)
Central and Southern MD (> 5 counties: Carroll, Howard, Anne Arundel, St.Mary's, Charles; ~2,300 sq.mi, Airborne-1 and Sanborn data, 2-m resolution)
Camden and Burlington Counties, NJ (1050 sq.mi, Airborne-1 data, 1-m intensity imagery)

Almost all projects include generation of DTM (usually with 2-m resolution) and intensity imagery (1-m resolution). Many projects also include 3D-modeling of buildings and linear features extraction. The map shows our MD projects during 2003-2006 for the MD Department of Natural Resources/Spatial Systems Associates and Federal Emergency Management Agency/Dewberry.

2010 Greenwich Institute for Science and Technology