Abstract

The presence of buried landmines is a serious threat in many areas around the World. Despite various techniques have been proposed in the literature to detect and recognize buried objects, automatic and easy to use systems providing accurate performance are still under research. Given the incredible results achieved by deep learning in many detection tasks, in this paper we propose a pipeline for buried landmine detection based on convolutional neural networks (CNNs) applied to ground-penetrating radar (GPR) images. The proposed algorithm is capable of recognizing whether a B-scan profile obtained from GPR acquisitions contains traces of buried mines. Validation of the presented system is carried out on real GPR acquisitions, albeit system training can be performed simply relying on synthetically generated data. Results show that it is possible to reach 95% of detection accuracy without training in real acquisition of landmine profiles. Document type: Conference object

Abstract

Traditional geotechnical investigation consisting of drilling many of boreholes, soil sampling and lab testing becomes uneconomical and time consuming for large and geologically intricate areas. In such cases, non-destructive geophysical surveys offer a time-efficient and economical solution, simultaneously providing an extensive areal coverage of subsurface profiling of geotechnical and geophysical properties. For the estimation of subsurface properties, three geophysical methods were employed in this study, which consisted of Ground Penetration Radar (GPR), Electrical Resistivity Tomography (ERT), and Multi-Channel Analysis of Surface Waves (MASW). Thus, the sub surface’s shear wave velocity profiles and apparent resistivity profiles were obtained. In the shallow depths, discontinuities and fractures were studied using electromagnetic radargrams from GPR. These tests were complemented with borehole exploration at the site. Results obtained from surface-based methods were validated against bore log data and visual observation of surface terrain. Subsequently, bore logs were integrated with the geophysical survey results to construct an integrated subsurface profile. The integrated subsurface profiles presented a three-layer subsurface structure consisting of dense gravelly sand in the top 1m, followed by strong rock formations and, ultimately, very strong fractured rocks. The extent of fracture in the rocks was studied using samples obtained from boreholes and available trenches at the test location. These findings helped compare the three methods and their applicability in delineating different subsurface layers in this study. GPR proves to be effective at shallow depths, while ERT and MASW help investigate deeper layers better. Further, this study offers critical insights for site characterization and engineering decisions in complex geological environments, improving the knowledge base of efficient and reliable subsurface evaluation techniques.

Abstract

Coastal beach and dune ecosystems are critically important for shoreline protection and significant resources have been allocated to their conservation. Dune vegetation is known to modify site response to wind, waves, and storms, but little focus has been given to understanding how belowground sediment structures enhance dune stability. A first step in addressing this knowledge gap is to determine optimal methods to measure subsurface sediment properties in onshore sandy environments. Our team performed a comparison of geotechnical and geophysical methods on a restored beach and dune system on Florida’s Atlantic coast. Methods included two types of dynamic cone penetrometer (DCP) systems, a multi-channel analysis of surface wave (MASW) system, a ground penetrating radar (GPR), soil vibracores, and grain size analyses. Two transects were investigated, a 50-meter cross-shore transect and an intersecting 20-meter along-shore transect. The transects were selected to study the gradient of sediment properties from the swash zone to the high dune. Key findings from this study were that the PANDA DCP provided a higher resolution of measurements compared to the standard DCP, which is extremely advantageous in the shallower, less consolidated soils. MASW shear wave velocity results showed similar trends to the DCP cone tip resistance and allowed for measurement of stiffer soils where the DCP reached refusal. In addition, the results from the DCP, GPR and vibracores compared well. Thus, the use of the DCP and GPR systems as minimally intrusive testing options in fragile dune systems was verified.