Floodplain land cover maps are important tools for river management and ecological assessment. These maps have to be revised regularly due to the dynamic nature of floodplains. Automation of the mapping process cuts the map processing time and can increase overall accuracy. Manual delineation and classification of vegetation is based on colour, contrasts (texture) and often-stereoscopic viewof aerial images. In this study, this is mimicked by the simultaneous use of both structural light detection and ranging (LiDAR) and compact airborne spectral imager (CASI) remote sensing data in an image segmentation routine fractal net evolution approach (FNEA). The segmentation results are tested against manually delineated ecotopes. Ecotope delineation improves when LiDAR and CASI are used simultaneously; the combination significantly lowers the number of segmented objects needed to accurately map ecotopes. Overall map accuracy of the LiDAR and CASI combination is 8–19% higher than single CASI and LiDAR, respectively.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below