Date of Award

Winter 2020

Project Type

Thesis

Program or Major

Earth Sciences

Degree Name

Master of Science

First Advisor

John E. Hughes Clarke

Second Advisor

Brian Calder

Third Advisor

Thomas C. Weber

Abstract

Because the safety of the navigation depends on accurate knowledge of the submerged features, any improvements in the ability to resolve those features are of major interest. Ultimately this reflects the bottom detection performance of the bathymetric measuring system (most commonly multibeam sonar) utilized. To this end, different algorithms for the detection of the seafloor or other targets have previously been developed, all presenting advantages and disadvantages. The two most common techniques are designed around time-series or angle-series analysis, although by far the most focus has been on time-series. The option of recording both amplitude and phase data of the water column permit the development and testing of new algorithms to be carried out in post-processing.This research evaluates the use of water column data to perform bottom detection in a nonconventional way, that is, analyzing angle-series instead of time-series. The proposed algorithm is based on the Beam Deviation Indicator (BDI) method with the inclusion of phase information. It sequentially evaluates each time-slice (angle-series), applying absolute and relative threshold filters to select echo envelopes based on intensity data. Then, the phase data of each echo envelope is analyzed for zero-crossings (across beams), which are then converted into angles. Thus, time-angle pairs are obtained, defining depth measurements. Such a method, herein termed Phase Deviation Indicator (PDI), can be applied in an alternative and mainly complementary way to the currently existing methods. Four different shallow waters seafloor relief types were investigated using an EM 2040P MkII multibeam echosounder, and collected datasets were evaluated with a focus on target or seafloor detection in many different geometries. The results obtained indicate that there are cases in which the analysis made only within a time-series, even using multi-detection features, can be incomplete and could be complemented by the analysis made within a beam-series (PDI). Particularly notable geometries included mast-like-objects, discontinuous surfaces or features whose lateral extent is confined mostly within a short range of incidence angles, thereby requiring multiple detections within the same beam. Such results emphasize the idea that the best detection method results from the integrated use of all available techniques.

Share

COinS