Tidal flats, megaripples and marsh: automated recognition on aerial images
Kleinhans, M.G.; Douma, H.; Addink, E.A.; Jentink, R. (2019). Tidal flats, megaripples and marsh: automated recognition on aerial images, in: Lefebvre, A. et al.International conference MARID VI - Marine and River Dune Dynamics 2019, 1 - 3 April 2019, Bremen, Germany: book of abstracts. pp. 135-140
Intertidal areas in estuaries are highly dynamic and ecologically important. Here we present a novel method for automated recognition of saltmarsh, pioneer vegetation, megaripples and plane bed from aerial photographs and bathymetry. Results are compared with manual classification. The two methods agree for about 73% of the whole area, and 95% for (partly) vegetated areas. The distinction between the bare bar surface classes is more difficult and often vague due to the existence of gradients and temporal transitions. The distinction between (nearly) plane sand bed and megaripples does not improve much with the use of slope and slope variation calculated from bathymetry, but further improvements may be possible with data derived from stereo photography. Nevertheless, the high degree of agreement shows that the automated method is highly promising for application to large systems such as the Wadden Sea.