Denoising Ground Penetrating Radar Images Using Generative Adversarial Networks

The objective of this study is to improve the visual quality of noisy ground penetrating radar (GPR) images by developing a deep learning network. The dataset includes noisy GPR images that were generated by adding white noise with a coefficient of variation of 1.0, and the original raw GPR images as labels. In addition, a denoising deep learning network was built and trained on the dataset. The experimental results show that the denoising network performs well with high peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) values. Furthermore, the denoised GPR images show as much detail as the ground-truth images. This study shows that the denoising network significantly denoises the GPR images.