Signal Processing Techniques
Richard V. Sailor · 2020
This chapter discusses smoothing and spectral analysis as applied to time-series of satellite altimeter data. Techniques are developed for analyzing the signal and noise characteristics of the data. A section on signal detection covers automatic detection of the geoid signatures of seamounts and fracture zones. The value of modeling the gravity field itself as a random process is well established in “statistical geodesy”, despite conceptual difficulties such as how to define ensembles and realizations (since there is only one earth) and the fact that the gravity field in any area is rarely stationary and isotropic. Signal processing techniques may be grouped into three main categories: filtering and smoothing, spectral analysis and detection and classification. Many of the most widely used signal processing techniques were originally developed for analog applications in electrical engineering, particularly for audio and telecommunications applications. These techniques are based on representations of signal and noise as continuous functions of time.