Image analysis for quantifying the spatiotemporal evolution of rill networks in laboratory experiments

This study describes technology developed to quantify surface flow width change at high temporal (1-min) and spatial (0.5-cm) resolutions. A series of laboratory experiments were performed to quantify the emergence, evolution, and resiliency of rill networks. An automated methodology was created to process and convert large number of raw data produced into actionable information. The proposed methodology combines image spectral transformations, morphological analysis, optimized regression, and GIS analysis to determine spectral discontinuities representing surface flow edges. Surface flow width measurements from the automated system were evaluated using three different variations of the algorithm. Results depicted agreement between these methods and revealed discrepancy between measurements generated by independent analysts due to human subjectivity. This methodology supports the development of modeling technology for improved estimation of rill and ephemeral gully network evolution and their sediment load production.