Detection and description of sand dunes using a geomorphometric approach
Résumé
In the last decades, sandbanks and sand dunes of various shapes and sizes have been found and
described on continental shelves all over the world. These structures tend to migrate and our
understanding of the mechanisms regulating their dynamics is limited. These sandbanks became a
concern due to the fact that they interfere with the human activities at sea. Consequently, their
presence has economical and environmental impacts. The purpose of this study is to automatically
detect these dunes in DTMs (Digital Terrain Models) generated from MBES data. So far, automatic
techniques aiming at describing marine dune fields have always brought information such as the
height, dune spacing or the orientation at local scale but never went down to a dune scale.
Currently, the dunes are still manually digitized and then form parameters are calculated. This task is
tedious and time-consuming. In addition, the amount of data to process being always bigger, it gets
difficult for operators to keep the pace. To address this issue, we propose a geomorphometric
approach. Attributes such as the slope or curvatures are estimated from DTMs. A combined analysis
of the metrics maps enables to extract dunes from their environment. The following step is the
determination of parameters characterizing the shape and size for each dune. This work provides
knowledge about dune fields at a detailed level. Thus, the composition, spatial distribution and other
aspects of sand banks can be obtained. Further research can be conducted in order to learn more
about the dynamics of dune fields at an individual level.