Sand-Spit Evolution and Inlet Dynamics Derived from Satellite Images: A Case Study for Tien Chau Inlet, Vietnam

Multispectral satellite imagery and machine learning for the extraction of shoreline indicators

“…A Multilayer Perceptron (MLP) segments the image into 4 classes allowing only the sand/water classes to be used when applying the MNDWI and Otsu's histogram thresholding to the image. >20m 10m -<20mTung et al (2021) The waterline is detected using, CoastSat (Vos et al 2019) with a MLP, the MNDWI with Otsu'were used to classify land and water classes including MLP, Artificial Neural Network (ANN), Decision Tree (DT), Naïve Bayes (NB), K-Nearest Neighbour (KNN), and SVM. >20m Minghelli et al (2020) SVM classification method was compared with four other methods the Euclidean Distance (ED), the Spectral Angle Mapper (SAM), MLC, SVM and ANN <10m Almonacid-Caballer et al (2016) The optimum threshold is obtained from a histogram of an infrared band when water and land are both present.…”