Combined low-rank and sparsity filtering of radio-sounding data

This paper presents a combined low-rank and sparsity filtering method to improve signal-to-noise ratio (SNR) of the radio-echo sounding (RES) data of ice sheets. It performs denoising in spectral/temporal subspace and spatial subspace based on low-rank approximation and sparse representation respectively. Experimental results show that the proposed method can denoise the RES data efficiently while preserves layer features well.