Author :
Olhoeft, G.R. ; Smith, Samuel ; Hyslip, J.P. ; Selig, E.T.
Abstract :
Over the past few years, we have collected over a terabyte of data along railroad tracks, developing a method to use ground penetrating radar to evaluate a variety of maintenance-of-way problems. An automated system to do this has been developed as the ultimate requirement is to perform such surveys twice per year over 400,000 km of track in the United States. At a data acquisition rate of 10 GBlkm, the data storage and processing requirements are enormous. With this quantity of data, most of it will never be seen by human eyes, so automation is essential. The operating requirements include high-speed GPR capability to image along and across the track, the ties (sleepers), the ballast and subgrade at railroad speeds (track time is limited and expensive). The surface of the railroad right of way has a variety of rough surface and volume scattering or clutter elements, including the rail and ties, sensors, cables, wavelength sized gravel ballast, debris, maintenance objects (spare ties and rails, sand barrels, etc.), power and comimunication systems (including third rails and overhead or underground systems) requiring multiple polarization, mull ichannel, multioffset GPR. Elevated air-launched horn antennas are required to clear switches, debris and sensors in the traclk. For their purposes, the railroads need to locate a problem area to about a meter, but to process and model GPR data adequately requires more spatial accuracy. DGPS doesn?????????t work in tunnels, is sometimes unreliable because of thunderstorms or geomagnetic storms, and deteriorates in EMI/RFI rich urban areas or where satellite coverage is limited, so multiple and redundant location positioning systems are required. We use SBASIWAAS DGPS for precise location positioning complemented by wheel encoder DMI or 6 DOF IGS in tunnels and other GPS unfavorable places, further backed up by redundant digital video, and quality tested by mapping against USGS 1-m DOQ´s. Integrated multichannel digital video- is also required to quality control the field of view of the radar antennas and to locate and inventory railroad assets and condition. Automated processing, calibration and modeling of the data through to simple maintenance-of-way indices useful to the railroads are being integrated into the railroad maintenance record systems along with archival data recovery procedures and data heritage traceability for accident investigations.