|
| | Abstract | | There is a high demand from administrators of buried objects for services to provide information on 3D underground pipe models with accurate position. The authors tried to automatically extract the position of the top surfaces of buried pipes from densely acquired GPR data with RTK-GNSS, and to model the buried pipes in a 3D-dxf file. We acquired GPR data in our two test sites for the detection of buried pipes. We scanned reflection points with determining velocity by maximizing semblance along the travel time curve of reflection wave under the assumption of a constant velocity and point reflection. We refer to the estimated velocity as apparent velocity. The extracted reflection points were filtered to select both points distributed on a straight line and points with the same apparent velocity, and points laying on a straight line were grouped. The depths of pipes were estimated by the correction coefficient of apparent velocities, which is the ratio of apparent velocity to true velocity in relation to the angle between a pipe and a survey line calculated based on a simple simulation. Finally, we built 3D solid models of buried pipes by giving the diameters of the pipes. Although we used a single channel antenna system which took much time to acquire spatially dense GPR data in the study, we plan to apply to GPR data efficiently acquired by multi-channel antenna system in future. |
|
|
|
| |
|