Summary
Seismic data are unavoidably contaminated with noise. The issue becomes worse for applications in mineral explorations in hardrock environments where the complex near- and sub-surface conditions result in variable signal amplitudes and attenuations (Eaton et al., 2003). Meanwhile, the quality of processing is usually uncontrollable, consequently, yielding remnant local noisy events as singularities on stacked seismic images. These singularities perform as backscattering sources during the migration process to create artificial features interfering with interpretation. No post-migration processing technique has yet demonstrated effectiveness to reduce this artificial noise. This problem can be solved by the curvelet transform for the efficiency of representing directional features on seismic images. We investigate the elimination of the migration-induced noise using a 2D fast discrete curvelet transform (FDCT) threshold applied to a synthetic Stolt time-migrated impulse response. The result shows a significant removal of migration-induced noise. This improvement is also applied to a 3D post-stack time-migrated (PoSTM) seismic cube, resulting in an enhancement of the original signal and a suppression of the induced noise.
Introduction
Hardrock seismic data acquired for mineral explorations often have low signal-to-noise ratio (SNR) and spurious (or discontinuous) reflection event. The migration of these low-quality images produces poor results by introducing backscattered "arcs" (Warner, 1986). In addition, problems such as that within the algorism, truncation effects, dispersion, ray bending and spatial aliasing intensify the weakness of migration application in hardrock environments (Milkereit, 1987). Curvelet transform, here we find, is an ideal tool to eliminate the migration induced noise, thus, is robust for improving image quality for seismic data in mining.
In 2007, a 3D vibrating-controlled seismic survey was performed at the Millennium mine site for uranium exploration in the Athabasca Basin, Saskatchewan, Canada (Juhojuntti et al. 2012, Wood et al. 2012). The foremost objective of the seismic survey was to map an unconformity at 500 to 700m depth underlying a metamorphosed sandstone sequence. The occurrence of the uranium ore deposit was confined by this high-contrast unconformity and fault associated hydrothermal alterations.
