Using K-Means and morphological segmentation for intertidal flats recognition

Tidal flats are considered an invaluable natural resource. Generally, they show dynamic morphologic changes that arise from high tidal energy and sediment transportation. The study of sediment budget processes is important in many ecological systems. The sediment budget can be estimated only if maps of morphological changes are available. Remote sensing, combined with in situ surveying, is an effective tool for monitoring tidal flats. The aim of this work is to precisely map tidal flats changes using multi-temporal Synthetic Aperture Radar (SAR) images acquired by the TerraSAR-X sensor. For that purpose, is proposed the application of K-Means clustering (in a gray level mode) to both original and preprocessed SAR amplitude data sets, followed by a morphological sequence task to generate optimized datasets, which will be compared for change detection between tides. Results will show the segmentation of intertidal regions from flooding tide´s comparison.