Depth-estimation and seabed-visualization techniques for airborne laser bathymetry.

Date

1995

Authors

Cheng, Wei

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Several techniques for the efficient processing of bathymetric data are developed. First, the characteristics of bathymetric signals are analyzed in both time domain and frequency domain. In the time-domain analysis, a bathymetric signal is charac­terized by three components that represent the blue-green surface reflection, volume backscatter, and bottom reflection. The characterization of the waveform facilitates decomposition of the waveform and renders the depth estimation reliable and phys­ically meaningful. Recursive computational schemes for the evaluation of the signal model are derived. Our simulations have demonstrated that these recursive schemes can be used in optimization-based depth estimation algorithms with reduced compu­tation complexity. In the frequency-domain analysis, the short-time discrete Fourier transform (STD FT) of the signal is used as a tool in the study of sea-depth estima­tion. In particular, we show that the STD FT evaluated at zero frequency provides an enhanced signal profile from which accurate estimates of sea depth can be obtained. In the second part of the thesis, noise filtering techniques for obtaining improved 2-D seabed topography of coastal shallow waters are investigated. The characteris­tics of the noise with which the bathymetric signals are contaminated are studied. A simple, yet efficient, method for regularizing irregularly spaced data is developed. Then a 2-D noise filtering algorithm based on the use of local statistics of depth measurements is described and applied to the data. The aforementioned processing involves only a small window of data surrounding the grid point being processed; consequently, the proposed algorithm can be implemented in real or quasi-real time.

Description

Keywords

Citation