Statistical Processing of Multibeam Sonar Data from SAX-99


The wide area coverage and great data density afforded by quantitative multibeam sonars provides the potential for statistical evaluation of seafloor backscatter and an understanding of the characteristics of the seafloor. The multibeam data gathered using a Simrad EM3000 sonar during SAX-99 is particularly suited for quantitative seafloor studies because of the tremendous overlap in the data gathered, the constant depth range of the seafloor, the availability of auxiliary data sets and the relatively homogeneous seafloor types in the survey area. A total of 307 lines were gathered in several directions with an average line offset of 10m in a central area of $600\times 600$m, resulting in an average density of 125 soundings/m${}^2$. The massive redundancy provides for robust analysis of the data and allows for error analysis, extraction of a statistically realistic representation of the bathymetry, and thus true radiometric correction of backscattered values. This geometry also allows us to evaluate mean backscatter as a function of grazing angle for any given piece of seafloor, thus obviating the need to assume a homogeneous seafloor across the swath. These preliminary analyses allow us to approach the problem of remote seafloor segmentation and characterization using the dependence of backscatter on grazing angle with more confidence.


Center for Coastal and Ocean Mapping

Publication Date


Journal or Conference Title

Fall Meeting, American Geophysical Union (AGU)


81, Issue 48

Conference Date

Dec 15 - Dec 19, 2000

Publisher Place

San Francisco, CA, USA


American Geophysical Union Publications

Document Type

Conference Proceeding