A ground penetrating radar denoising method based on CEEMD and wavelet decomposition

Abstract

Ground penetrating radar (GPR) technology is widely used in civil engineering projects such as inspection of concrete buildings, pavement road, bridge, tunnel, and underground utilities (water supply pipes, gas pipes, power cables, sewers, etc.). With the application of GPR becoming more and more extensive, in order to further improve its effectiveness and efficiency, the researches on noise removal, image quality improvement and automatic data interpretation are being actively carried out. In this paper, firstly, the principles of complementary ensemble empirical mode decomposition (CEEMD) and wavelet transform, which are widely used in various signal processing fields, are described in detail. Then, GPR denoising method based on CEEMD and wavelet decomposition is proposed. The CEEMD IMFs of GPR signal separate into the effective signal components and noise ones. The noise components are not completely removed, the effective information are extracted from them by the wavelet denoising technique and then the signal is reconstructed. Finally, the performance of proposed method is verified by numerical simulation and field data analysis.