*Important notice: This news reports on an unedited version of the paper which has been accepted. and is awaiting final editing. Scientific Reports sometimes publishes preliminary scientific reports that are not fully edited and, therefore, should not be regarded as conclusive or treated as established information.
High-resolution optics methods using time-reversal and MUSIC processing improve radar imaging accuracy. This optics approach enhances underground object detection and reduces noise in subsurface mapping.
Study: Enhanced GPR imaging using high-resolution TR-MUSIC for underground object localization. Image Credit: vectorfusionart/Shutterstock
A new study in Scientific Reports presents a high-resolution imaging approach for time-reversal focusing with subspace-based Multiple Signal Classification (MUSIC) processing to improve underground object localization. The researchers show that the method can be applied directly to conventional systems without requiring additional hardware or complex measurement setups.
The study demonstrates that the proposed approach overcomes key challenges with simulations and experiments confirming significantly sharper imaging and improved detection accuracy compared to traditional techniques.
Advancing Ground-Penetrating Radar Imaging
Ground-penetrating radar (GPR) plays a central role in detecting and mapping subsurface structures in civil engineering, archaeology, and environmental monitoring. Strong surface reflections and surrounding clutter often mask signals from buried objects, particularly those located near the surface. Detecting deeper targets requires lower-frequency signals, but this reduces spatial resolution and makes fine structural details harder to resolve.
Researchers have explored advanced signal processing techniques such as time-reversal (TR) and time-reversal multiple signal classification (TR-MUSIC) to address these challenges. These methods enhance focusing and resolution by leveraging wave propagation and scattering behavior.
However, they remain difficult to implement in practice. Conventional TR-MUSIC depends on multiple antennas and accurate knowledge of subsurface conditions, and even small mismatches between models and real environments can reduce performance.
This study aims to retain the high-resolution advantages of TR-MUSIC while removing its practical constraints. The proposed High-Resolution Time-Reversal (HRTR) technique processes conventional GPR data directly and operates with simple antenna configurations.
The technique avoids the need for antenna arrays and eliminates complex forward–backward modeling. This approach bridges the gap between high-performance theoretical methods and practical GPR applications, offering a solution that is both accurate and easy to implement in real-world settings.
Methodological Framework and Computational Approach
The HRTR method integrates time-reversal signal processing with the MUSIC algorithm to achieve enhanced subsurface imaging. The workflow begins with the acquisition of GPR signals in either monostatic or bistatic configurations. These signals, which contain reflections from subsurface structures, are transformed into the frequency domain to construct the system’s transfer function.
The method then applies time-reversal processing by reversing the received signals and re-injecting them into the medium. This step refocuses electromagnetic energy toward the original scattering sources, improving localization compared to conventional approaches.
HRTR incorporates subspace-based processing by constructing a time-reversal operator and applying singular value decomposition (SVD) to separate the data into signal and noise subspaces. The signal subspace captures meaningful reflections from subsurface features, while the noise subspace represents background clutter and interference.
The MUSIC algorithm is subsequently implemented in the time domain using delay-based steering vectors. This enables precise identification of reflection times associated with subsurface discontinuities. The resulting pseudospectrum exhibits sharp peaks that directly indicate object locations.
The researchers validate the method through numerical simulations using the gprMax electromagnetic solver, covering both two-dimensional and three-dimensional scenarios with varied soil conditions, object geometries, and excitation frequencies. They further confirm its performance through laboratory experiments and field measurements using standard GPR equipment, including antennas and a vector network analyzer, demonstrating the method’s practical applicability.
Enhanced Imaging Performance: Results and Interpretation
The results demonstrate significant benefits in both imaging resolution and object localization. In simulation studies, the HRTR method produces sharper and more confined reflections than conventional radargrams. Traditional GPR images often appear blurred and temporally spread, whereas HRTR generates narrow, well-defined peaks that accurately mark object positions.
Conventional GPR often merges surface and object responses into a single signal when objects are located close to the ground surface. HRTR resolves these components distinctly, highlighting its superior temporal resolution. This capability is especially valuable in critical applications such as landmine detection.
The method also shows strong performance across different soil conditions. In low-loss environments, it delivers precise and high-contrast localization. As soil conductivity increases, signal attenuation reduces overall performance, but HRTR continues to outperform conventional techniques until attenuation becomes dominant. HRTR achieves high-resolution imaging using significantly fewer frequency samples than traditional methods.
The results further indicate that using either the signal or noise subspace yields nearly identical outcomes, highlighting the robustness of the approach. Three-dimensional simulations confirm these benefits under more realistic conditions, including varying object orientations and polarization effects. In all cases, HRTR produces clearer images with reduced clutter.
Download the PDF of this page here
HRTR delivers sharper imaging and improved separation between targets in the laboratory tests with buried metallic objects. Field experiments further demonstrate its effectiveness, successfully detecting multiple buried objects, including landmines. HRTR reveals fine structural details, such as internal features of a copper ring, that remain indistinct in conventional GPR images.
Conclusion and Broader Impact
The HRTR method marks a significant step forward in GPR imaging. By combining time-reversal focusing with subspace-based analysis, it delivers high-resolution results while remaining fully compatible with standard GPR systems. The approach removes the need for antenna arrays and avoids complex modeling, making it well-suited for practical, real-world deployment.
The study shows that HRTR improves both resolution and detection accuracy. It clearly resolves closely spaced objects and enhances the detection of shallow targets. Additionally, it supports the use of low-frequency signals for deeper penetration without compromising resolution. This balanced performance addresses a long-standing limitation in conventional GPR systems. The method can support applications in infrastructure inspection, environmental monitoring, archaeology, and humanitarian demining.
The researchers provide a graphical user interface and open-source code, enabling researchers and practitioners to apply HRTR directly to existing datasets. This practical design encourages wider adoption and further innovation.
Future work should focus on optimizing the selection of singular values in the subspace decomposition, which could enhance performance and enable more automated processing. Overall, HRTR closes the gap between advanced signal processing theory and real-world GPR applications.
Journal Reference
Karami, H., Romero, C., et al. (2026). Enhanced GPR imaging using high-resolution TR-MUSIC for underground object localization. Scientific Reports. DOI: 10.1038/S41598-026-49191-X, https://www.nature.com/articles/s41598-026-49191-x
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.