The Application Of Image Processing And Analysis Algorithms To Complex Gpr Datasets

In suitable environments, ground penetrating radar (GPR) offers the highest resolution representation of the shallow subsurface available in the geophysical toolbox. Potential applications of<br>the technology have burgeoned over the last 30 years from academic and archeological pursuits by a handful of researchers to the myriad of industries and environments of today. Regardless of the current commonality of GPR, the approaches used for data processing and interpretation, based primarily on reflection seismics, have persisted. Although the analogy between acoustic and electromagnetic wave reflection profiling is valid, GPR can yield raw data which are orders of magnitude richer in detail. Recent advents in rapid sampling circuitry and large multi-channel arrays have rendered the traditional approach of manually discerning point reflectors or horizons insufficient. Rather than being relegated to the onfines of seismic reflection processing, this paper explores a novel approach to the processing and analysis of GPR data by considering radar profiles as raster images. Image processing methods, commonly employed in microscopy, industrial quality control, forensics and radiology, offer the ability to extract subtle textural or reflector orientation variations, enhance or subdue various temporal or spatial frequencies, and automatically detect targets based on image morphology. This approach to data processing may offer GPR researchers and practitioners the ability to extract dramatically more detailed and consistent information from profiles which were previously under-interpreted.