Accurate gain method for ground-penetrating radar signals based on stationary wavelet packet transform

Abstract

In this study, we propose an accurate gain method for ground-penetrating radar (GPR) signals based on the characteristics of refined time-frequency analysis and translation invariance offered by the Stationary Wavelet Packet Transform (SWPT), combined with the conventional signal gain approach. This method aims to address the issue of low signal resolution resulting from the direct gain processing of GPR signals with a low signal-to-noise ratio (SNR). Specifically, the GPR signals are initially decomposed into appropriate wavelet packet coefficients using SWPT, wherein only those coefficients with high SNR undergo gain processing, followed by reconstruction of the signals through SWPT. By employing accurate gain processing on low SNR GPR signals acquired during concrete crack detection tests, we have confirmed that the proposed method effectively distinguishes the target reflected signals from most noise, thereby achieving accurate amplification of the desired reflected signals and significantly enhancing the GPR signals resolution under low SNR conditions.