Precise Airborne LiDAR Surveying For Coastal Research And Ge(3)

International Archives of Photogrammetry and Remote Sensing, Volume XXXIV-3/W4 Annapolis, MD, 22-24 Oct. 2001

Figure 4. The Galveston Bay - Bolivar Peninsula area.

Figure 5. Baseline distance during 17 July shoreline survey.

Figure 6. Difference in HAE between the Port Bolivar andSabine Pass aircraft trajectories for July 17, 2001 shorelinesurvey.

We computed KARS trajectories for the aircraft using both thePort Bolivar and Sabine Pass GPS base station data. Thedifferences between the two trajectories in the east and north,components are under 0.05m. The HAE differences between thetwo trajectories are under 0.05m when the aircraft is within50km of both base stations. The HAE differences are under0.10m even when the aircraft is more than 100 km from one ofthe base stations (see figure 6).2.3 Data Coverage

Small foot-print ALSM systems operating with a 25kHz orhigher laser pulse repetition rate can generate ALSM coverage

187

with a sub-meter laser point spacing during a single pass.However, vegetation, buildings, and topography can causeshadowing that may significantly reduce the ground surfacecoverage. For area surveys we fly an orthogonal grid, two sets offlightlines at right angles, to minimize data gaps. Scanning from anumber of different aircraft positions allows us to more accuratelyreconstruct the morphology of topographic or cultural features. ForALSM surveys that are route-oriented, e.g. a shoreline survey,parallel swaths can be spaced laterally so as to scan both sides of aroute-parallel obstruction such as a dune line.2.5 Ground Truth

We conduct ground GPS surveys within each ALSM survey areato acquire ground “truth” information. We re-occupy the ALSMGPS base stations and survey an open area with an unambiguoussurface (road, soccer fields, large building) using kinematic GPStechniques. The ALSM data are sorted to find LIDAR points thatfall within 0.5m of a ground GPS survey point. The meanelevation difference between the ALSM (last returns only) and theground GPS are used to estimate and remove an elevation biasfrom the ALSM. The standard deviation of the elevationdifferences provide an estimate of the LIDAR precision. Selectedportions from each ALSM data set (last return only) are used togenerate a high-resolution (1m × 1m or 0.5m × 0.5m) digitalelevation model (DEM) or laser intensity image. The kinematicGPS data are superimposed on the DEM or intensity image andexamined for any horizontal mismatch.

Figure 7 is a 0.5m × 0.5m laser backscatter intensity image of thesoccer field in Juticalpa, Honduras. The chalk markings on thefield are discernible. On the right panel, the survey points from aGPS survey of the chalk marks and two transects across the fieldare superimposed on the intensity image. The GPS and ALSMmatch to within the resolution of the image indicating an ALSMhorizontal error of <0.5m. There were 417 ALSM points that fellwithin 0.5m of a GPS ground survey point on the soccer field. Themean elevation difference between the GPS and ALSM was –

0.169m with a RMS of 0.088m.

Figure 7. Intensity image of soccer field with GPS ground surveyoverlain.

2.4 Data classification

ALSM generates a semi-random cloud of elevation points thatrequires classification into reflections from ground and vegetation.As a preliminary step towards constructing digital elevation