GPR often encounters difficulty in visualizing the buried target when signals are weak and enveloped by noise, despite using the best of existing GPR data analysis tools. In this study, new method has been proposed based on modified S-transform to analyze weak signals of GPR data. The time-frequency analysis has been implemented to capture the change in electromagnetic waves reflected at the boundary of dielectric contrast. A series of experimental tests were conducted at different site conditions and with different materials using GSSI SIR 3000 GPR system with 400 MHz antenna. An improved 2-D vertical subsurface profile has been developed using time-frequency information to locate the target spatially as well as temporally. The new method has provided better results to improve the target visualization in case of noisy signals. However, it did not add any significant where conventional processing could do reasonable. The proposed method is convenient to use and capable of processing large number of scans effectively.

A new area at an existing archaeological site of Harappan civilization at Dholavira, Gujarat, India has been studied using ground penetrating radar (GPR). An area of 12,276 m 2 was surveyed using 200 MHz antenna at grid spacing of 2-3 m. The soil strata was found to extend mainly up to 3.5-4 m. The survey was conducted during the dry season to collect good signals. Post-processing was carried out to map the bedrock as well as archaeological features. A number of linear features were observed from the 3D image of the subsurface created from the acquired GPR profiles. Unlike residential structures, the large dimensions of these features indicate the likely existence of a series of water structures that may have partly collapsed due to floods at some point. There were some areas full of rubble next to the damaged walls that appeared to be orthogonal to the direction of possible flood from Manhar River.

No abstract

No abstract