Non-uniform Optimal Sampling for Seismic Survey Design

Defining optimal locations for sensors in a seismic survey has been a long-standing issue for geophysicists. It has long been known that random sampling can in some cases recover broader bandwidth from a fixed set of samples than uniform sampling. As an improvement over random sampling, we propose a methodology for Non-Uniform Optimal Sampling (NUOS) as a means for choosing source and receiver locations for seismic survey planning. This technique uses concepts from the field of compressive sensing in conjunction with optimization algorithms to identify sensor layouts that satisfy optimization constraints for a particular survey. Field trials conducted using NUOS concepts confirms the viability of using compressive sensing algorithms to recover significantly broader spatial bandwidth from non-uniform sampling than could be obtained using uniform sampling. Field trials using NUOS designs show that spatial bandwidth equivalent to 12.5m uniform sampling can obtained using the same number of samples as would be used for a 25m survey, resulting in significant improvements over conventional surveys. This improvement could be used to reduce costs for a fixed area, to cover a larger area with the same amount of equipment, or to increase the resolution over a given area.