A comprehensive study of some features from characteristics of enhanced ground‐penetrating radar wave images through convenient data processing within carbonate rock, west of Assiut, Egypt

Abstract

The Eocene limestone plateau, west of Assiut, is characterized by different sedimentary structural features such as fractures, karst, cavities and marble. All conventional methods failed to detect these features due to their inhomogeneities. In this study, a ground‐penetrating radar survey was applied to an area including an old quarry of marble lying at the eastern part of the limestone plateau. The field survey was carried out using a 200‐MHz antenna following the profiling and areal mapping techniques. The acquired ground‐penetrating radar wave images were processed and interpreted using the available software program to enhance the different reflection characteristics from the subsurface anomalous materials. A series of the three‐dimensional model cross section with variable depths were constructed to identify and characterize different georadar facies changes vertically and laterally. The obtained results were adjusted and compared with the excavated features of the quarry; therefore, ground‐penetrating radar anomalies and reflections were calibrated. In the present study, the ground‐penetrating radar survey succeeded in differentiating between dense limestone rocks, onyx marble and microstructural features resulting from recent tectonism. Moreover, the investigation of ground‐penetrating radar reflection waves is useful in differentiating between low‐ and high‐quality limestone and onyx marble. This study identifies the various features of the limestone plateau lying below the ground surface, west of Assiut, that are to be taken into account by planners when establishing new urban centres on the plateau in the future.

Rock Layer Classification and Identification in Ground-Penetrating Radar via Machine Learning

“…Furthermore, deep learning can handle large-scale data, speeding up the analysis process and increasing identification accuracy. In this paper, we have employed advanced deep learning techniques to harness GPR data for the identification of underground rock layers and geological strata, fully capitalizing on the remarkable capabilities of this technology [28]. We have utilized convolutional neural networks (CNNs), a powerful tool in this context, to accurately extract information pertaining to rock layers and strata through the analysis and processing of GPR images.…”