RiverDetectWood: A tool for automatic classification and quantification of river wood in river systems using aerial imagery

Abstract

The RiverDetectWood tool automates the classification and characterization of wood in river systems using very high spatial resolution (VHSR) aerial photographs. River wood plays a dual role in river management: it contributes to habitat diversity and geomorphological changes but can also increase flood risk and infrastructure damage. Traditional field-based river wood surveys are time-consuming and geographically limited. By integrating machine learning techniques and remote sensing, RiverDetectWood offers a cost-effective, scalable solution for monitoring river wood presence and extracting key variables, such as length, diameter, area, and volume. This tool is designed for easy use by river managers and researchers, facilitating long-term monitoring and decision-making.