Determination of Water Depth in Ports Using Satellite Data Based on Machine Learning Algorithms

One of the fundamental maintenance tasks ports is periodic dredging them. This necessary to guarantee a minimum draft that will enable ships access safely. The determination bathymetries instrument determines need for and permits an analysis behavior port bottom over time, in order achieve adequate water depth. Satellite data processing predict environmental parameters used increasingly. Based on satellite using different machine learning algorithm techniques, this study has sought estimate seabed ports, taking into account fact areas are strongly anthropized areas. algorithms were Support Vector Machine (SVM), Random Forest (RF) Multi-Adaptive Regression Splines (MARS). was carried out Candás Luarca Principality Asturias. In validate results obtained, acquired situ by single beam provided. show type methodology can be coastal bathymetry. However, when deciding which system best, priority given simplicity robustness. SVM RF outperform those MARS. performs better with mean absolute error (MAE) 0.27 cm, whereas 0.37 cm. It suggested approach suitable as simpler more cost-effective rough resolution alternative, estimating depth turbid than single-beam sonar, labor-intensive polluting.