Permafrost Subgrade Condition Assessment Using Extrapolation by Deterministic Deconvolution on Multifrequency GPR Data Acquired Along the Qinghai-Tibet Railway

The Qinghai-Tibet railway (QTR) stretches 1956 km on Tibetan Plateau, the “roof of the world” from Xining City, Qinghai to the City of Lhasa, Tibet, China. About one half (~ 960 km) of the total length of this railway is on permafrost subgrades. Frozen earth hazard and other subgrade problems lead to subgrade instability and affect the normal operation of railway transportation. It is critical to develop a rapid and efficient technique for railway foundation defect detection. Multifrequency ground-penetrating radar (GPR) provides a good tradeoff between imaging depth and resolution for railway subgrades inspection. We have developed a signal fusion method to extrapolate the higher frequency, higher resolution GPR signal into a greater depth based on the lower frequency, greater penetration signal using the extrapolation with deterministic deconvolution (EDD) algorithm. This paper first introduces the principles of EDD and demonstrates the procedure by applying it to a synthetic data set. Next, data preprocessing and filtering are discussed to prepare the field GPR data acquired on QTR to be suitable for carrying out EDD. Finally, the multifrequency radar signal is fused based on EDD to reach a higher resolution and signal-to-noise ratio for the fused radargram profile. Examples of application of this proposed approach at three sections provide a demonstration for characterizing the status of subgrade permafrost and recognizing subgrade problem in different depths beneath the QTR. Our field tests of EDD to multifrequency GPR data demonstrate that it is a promising technique for signal enhancement.