Tree Roots GPR Detection Based on 3-D Time-Reversal Signal Processing

In this work, we analyze the performance of the time-reversal (TR) technique applied to tree root mapping using ground penetrating radar (GPR) to improve the detection of these subsurface targets. TR techniques exploit the invariance of the wave equation under time reversal and can yield features such as superresolution and statistical stability. The TR method was applied to 3-D GPR data collected around an urban tree and compared with traditional signal processing steps. It was possible to determine the root positions and depths with good precision, as well as differentiate anomalies from roots that are very close to each other. All the mapped roots were confirmed by excavating them at the site. The proposed method has the potential value for the tree roots GPR detection.