Ice velocity is a critical parameter in river ice monitoring, playing a vital role in analyzing ice blockages, the freeze-up process, and ice jam locations. It provides essential references for freeze-up forecasting. This study utilized UAV (Unmanned Aerial Vehicle) remote sensing imagery from December 2023, during the ice run period, combined with the Pyramidal Lucas-Kanade Optical Flow Algorithm to quantify ice velocity distribution across multiple bends during the Yellow River's ice run period. The study identified low ice velocity cross-sections at bend apexes, where velocity decreased sharply, leading to ice floe accumulation and an increased probability of ice jams. Significant spatial variations in ice velocity within bends were observed. Curvature and constriction ratios were identified as key factors influencing ice velocity distribution. Bends with high curvature and significant constriction should be prioritized for monitoring and early warning. The study also revealed a correlation between the low ice velocity cross-section at the apex of Shisifenzi Bend and the location of the initial ice jam during the ice run period, suggesting its potential as a scientific guide for ice jam risk forecasting. These findings provide valuable insights for freeze-up forecasting and flood prevention in the Yellow River.
Keywords: Bends; Ice velocity; Pyramidal Lucas-Kanade Optical Flow Algorithm; The Yellow River.
© 2025. The Author(s).