HydroBound-ML 2026-03_Magnuszew dataset

<p dir="ltr">This dataset provides the comprehensive source data, model files, and output results developed for the HydroBound-ML framework. The primary remote sensing data was acquired using the Riegl VQ-880-GII topo-bathymetric system integrated with PhaseOne iXM-RS150F optical sensors (RGB & NIR) over a complex fluvial environment (Pilica and Vistula rivers, Poland).</p><p dir="ltr">💻 <b>Software Repository:</b> The complete Python application (HydroBound-ML Assistant v.1.0.7), source code, and documentation are openly available on GitHub: <a href="github.com/PWroblewskiWAT/HydroBound-ML" target="_blank" rel="noreferrer">github.com/PWroblewskiWAT/HydroBound-ML</a>.</p><p dir="ltr">The repository is structured into three main components to ensure full workflow reproducibility:</p><p dir="ltr"><b>1.

INPUT DATA</b></p><ul><li>Ortho RGB imagery (.tif)</li><li>Ortho NIR imagery (.tif)</li><li>NDVI spectral index raster (.tif)</li><li>NDWI spectral index raster (.tif)</li><li>NGRDI spectral index raster (.tif)</li><li>Project metadata file (.json)</li><li>Digital Terrain Model (DTM) with ellipsoidal heights (.tif)</li><li>Area Of Interest (AOI) polygon (.kml)</li></ul><p dir="ltr"><b>2. MODEL FILES (Updated - v2)</b></p><ul><li>HydroBound-ML model v2 weights (<code>HydroBound-ML_v2.joblib</code>)</li><li>HydroBound-ML v2 manual samples database (<code>HydroBound-ML_v2_manual_samples.joblib</code>)</li><li>HydroBound-ML v2 model training metrics (<code>HydroBound-ML_v2_metrics_log.txt</code>) <i>(