Velocity Estimation in Tidal Rivers Using a Combined Rule-based Feature Extraction and Feature Tracking Approach

Accurately estimating riverine currents is a formidable task, requiring high-fidelity spatial and temporal observations. In this study, we present an integrated methodology for the determination of currents in tidal rivers and estuaries combining MCC with well-established image processing techniques for feature extraction and tracking, based on neural networks and color-opponent vision. A novel element of this work is the application of a rule based approach for the separation of Lagrangian from non-Lagrangian features. Compared to output non-hydrostatic barotropic model rule-based feature extractors provide a better estimate of current extent, magnitude and direction than the direct application on the MCC method to single band imagery.