Multi-frequency GPR signal fusion using forward and inverse S-transform for detecting railway subgrade defects

Railway subgrade defect is one of the most important inducements that lead to potential hazards in railway transportation. Railway subgrade is a layered structure including the sleepers, the ballast, back-filled basement and natural formation. There are numerous different types of subgrade defects in each layer. Multi-frequency ground penetrating radar (GPR) is an effective tool for detecting railway subgrade defects. With the use of multi-frequency GPR data acquired at the same location along railways it is highly desirable to get high resolution and deep penetration for radar imaging with data fusion techniques. In this paper, we report the analysis result using the forward and inverse S-transform to a dual-frequency data set containing the 100- and 400-MHz GPR profiles acquired on the Qinghai-Tibet Railway using a special GPR system. In preparing the data for S-transform a series of pre-processing stages including resampling, amplitude balance is carried out. The dual-frequency GPR spectrums are first fused together through forward S-transform and expressed as a spectrum with broader bandwidth in frequency-domain. Then, the synthesized spectra are converted back to time-domain shown as the synthesized GPR time sequence with the inverse S-transform. The results show that the fused radar signal has higher resolution and deeper penetration for the purpose of detecting subgrade defects located in different depth with different size. The characterization conducted in this manner serves as an important guidance for subgrade defects counter-measuring treatment.