HR: 0830h
AN: G11A-0238    [PDF]
TI: Correcting for GPS Multipath Error in LIDAR Surveys Using Crossover Analysis
AU: * Borsa, A A
EM: aborsa@ucsd.edu
AF: Institute of Geophysics and Planetary Physics, University of California, San Diego, La Jolla, CA 92093-0115 United States
AU: Bills, B G
EM: bbills@ucsd.edu
AF: NASA, Goddard Space Flight Center, Geodynamics Branch, Greenbelt, MD 20771 United States
AU: Fricker, H A
EM: hafricker@ucsd.edu
AF: Institute of Geophysics and Planetary Physics, University of California, San Diego, La Jolla, CA 92093-0115 United States
AU: Minster, J B
EM: jbminster@ucsd.edu
AF: Institute of Geophysics and Planetary Physics, University of California, San Diego, La Jolla, CA 92093-0115 United States
AB: The quality of the range measurement from an airborne Light Detection and Ranging (LIDAR) survey is largely dependent on the accuracy of the GPS trajectory for the aircraft. GPS elevation error - which today is largely due to multipath effects at the aircraft and the GPS base station - contributes a major portion of the LIDAR vertical error budget. The usual practice of quoting an RMS value for the GPS component of the error budget implies that GPS noise is Gaussian, yet the true nature of the noise signal is time-varying with significant power at long periods. GPS noise with a 3-cm RMS can easily have more than 10 cm of total variability on a time scale of tens of minutes to several hours. We show examples from an airborne LIDAR survey over the open-pit Hector Mine where repeated flyovers of an area used for ground truth revealed large elevation biases between passes that could not be resolved by adjusting the (non-GPS) parameters of the LIDAR system. As part of the post-processing of a large kinematic GPS survey of the salar de Uyuni, Bolivia, we have developed an algorithm to correct time-varying GPS error using elevation mismatches at crossovers between vehicle paths. The survey was originally designed to incorporate a large number of crossovers for the purpose of determining survey repeatability, and we were later able to exploit the crossover difference observations to solve for a model of the actual error signal generating those differences. We give results from tests with synthetic noise and topography data indicating that this method removes more than two-thirds of the added noise from the topographic signal, and we show the excellent results obtained for the salar de Uyuni survey data. We believe that airborne LIDAR surveys incorporating crossovers at regular intervals can also benefit from the application of this algorithm.
DE: 1204 Control surveys
DE: 1294 Instruments and techniques
DE: 9820 Techniques applicable in three or more fields
SC: Geodesy [G]
MN: 2003 Fall Meeting