Autoclassification versus Cognitive Interpretation of Digital Bathymetric Data in Terms of Geomorphological Features for Seafloor Characterization

摘要

The determination of seafloor geomorphological features has always been a difficult task, and it was not until the advent of marine remote sensing techniques that seafloor features could be accurately discerned. Airborne acquisition of digital bathymetric data provides a wealth of information that can be interpreted in different ways. This paper considers the pros and cons of computerized autoclassifications versus cognitive interpretations of seafloor features. The continental shelf off the southeast Florida coast contains LADS (laser airborne depth sounding) surveys that are here used to compare and contrast automated classifications of bathymetry with cognitive differentiation of marine geomorphological features. There are advantages and disadvantages associated with each approach, and the choice of methods depends on the purpose or goals of the project. Once seafloor features have been cognitively discerned from enhanced, color ramped, and vertically exaggerated bathymetry, machine classifications can be compared with known units. Using ArcGIs (R) ArcMap software, five- and seven-class unsupervised isocluster autoclassifications were found to moderately represent known bottom topography, whereas the interactive supervised autoclassification closely approximated cognitively discerned bathymetric patterns. Hand-drawn or digitized cognitively derived maps were more generalized than supervised computerized classifications based on training fields. Overall, both methods were found to be beneficial approaches, as they complement each other.