groundfiltering

Ground filtering, sometimes written as groundfiltering, is a set of signal processing techniques used to remove or suppress signals originating from the ground or environment that contaminate measurements. The goal is to enhance signals of interest, such as subsurface reflections in geophysical surveys or target acoustic signals in recordings.

In seismology and geophysics, ground filtering often targets coherent ground noise such as ground roll, cultural noise, or atmospheric disturbances. In audio and structural monitoring, it removes mains hum, mechanical vibrations, or transfer vibrations fed through fixtures.

Techniques include frequency-domain filtering (high-pass to remove low-frequency tilt, low-pass for high-frequency noise, notch filters for specific lines like 50/60 Hz); adaptive noise cancellation using reference sensors; multichannel or spatial filtering such as array processing and f-k filtering to suppress waves arriving from known directions; and time-domain methods like Wiener filtering and wavelet denoising; deconvolution to remove instrument response; detrending and normalization.

Filters can introduce artifacts or phase distortion; non-causal filters can alter timing; avoid over-filtering which can remove real signal; edge effects; ensure sampling rate is sufficient to capture signal bandwidth; use validation with synthetic data or cross-checks.

Common applications include seismic data processing to enhance reflections; vibration isolation in structural health monitoring; audio restoration to remove hum; and environmental monitoring to separate sensor noise from legitimate signals.