Precise Airborne LiDAR Surveying For Coastal Research And Ge(2)

and pitch bias corrections, a scanner scale correction, and atiming correction for each TIM. These corrections were initiallymeasured in the manufacturer's laboratory facility and refined byflight testing. In the laboratory, range corrections were alsotabulated for varying intensities of laser backscatter. We re-estimate the instrument calibration by flight-testing before andafter an ALSM survey. Estimating GPS datum or ranging errorsrequires flying the instrument against "ground truth" - an area(e.g. road or airport runway) surveyed by ground GPS orconventional means. However, the scanner roll, pitch and scalebiases can be accurately estimated through the careful

comparison of overlapping flightlines (Burman, 2000).

Figure 1

. Laser backscatter intensity image of calibration area.

Figure 2. Roll and scale errors before and after adjustment.Figure 1 is a laser backscatter intensity image constructed fromseveral flightlines on the Texas coast. Indicated on the image isa kinematic GPS ground survey on a paved road oriented normalto the direction of four crossing flightlines. Figure 2 shows theelevation differences (+) between the ground GPS and one ofthese crossing ALSM flightlines processed using nominalcalibration settings. We estimated calibration corrections fromfour flights spaced over two weeks (July 12 through July 27,2001) of surveying. Plotted for comparison are the elevationdifferences (Ο) between the ground GPS and the same flightlineafter calibration adjustment. The consistency of the fourcalibration flights indicates that the ALSM system’s pointing

186

accuracy has a RMS of ≤0.01° and a scanner scale RMS of≤0.0006.2.2 GPS

The absolute positioning of the ALSM platform comes from GPS.Therefore planning the GPS component of the ALSM survey,operating the air and ground GPS equipment, and estimating theaircraft trajectory from the GPS observations are critical steps. Weconduct ALSM surveys during periods when the Dilution ofPrecision (DOP) is ≤3.5 as estimated for a 15° elevation mask. Weoccupy ground GPS base stations that have an unobstructed sky-view down to 10°-to-15° above the horizon and are free of RFinterference or significant multi-pathing. We use dual-frequency,12-channel GPS receivers in the aircraft (Ashtech Z-12) and on theground (Ashtech Z-12 or Trimble 4000SSi) to record data at 1Hz.The ground receivers use Dorne & Margolin chokering antennas toreduce multi-pathing and a Dorne & Margolin C146-2-1 antenna ismounted in the aircraft. All antennas have been calibrated by theNational Geodetic Survey’s (NGS) Geosciences ResearchDivision. The NGS measures the antenna’s L1 and L2 phase centervariations as a function of GPS satellite elevation (see figure 3).Unless our GPS observations are corrected for these phase centervariations, errors as large as a decimeter can be introduced into theheight component of the aircraft trajectory.

54

)

cm3( no2Dorne & Margolin C146-2-1 antenna

ita1L1 Phase Centerivar0L2 Phase Cente

C P-1-2-3

10

20

30

40

50

60

70

80

90

GPS satellite elevation (degrees)

Figure 3. Phase center error as a function of satellite elevation forthe C146-2-1 antenna.

We use the NGS’s kinematic GPS processing software, KARS(Mader, 1992), to estimate a double-differenced, ionospherically-corrected (L3), ambiguity-fixed, phase solution for the aircrafttrajectory. We use precise GPS ephemerides, computed by theInternational GPS Service (IGS) or the NGS, instead of thebroadcast orbits in the trajectory solution.

On July 17, 2001, we mapped the Texas shoreline from SabinePass to Galveston Island (see figure 4). A Trimble 4000SSireceiver occupied a tide gauge benchmark at Sabine Pass and anAshtech Z-12 occupied a tide gauge benchmark at Port Bolivar.During the almost three-hour survey, the aircraft was alwayswithin 50 km of one GPS base station, but could be as far as 150km from the other basestation (see figure 5).