Inversion of intertidal zone topography based on optimized random forest regression characteristic parameters

It is a fundamental task to monitor the topography and understand changes of intertidal zone for rational utilization sustainable development. A new method proposed identifying terrain zone, using ICESat-2 data replace large amount on-site observation data, thereby reducing costs improving efficiency. Based on pre-experiments correlation analysis, time phase index, water transparency index suspended sediment concentration are added as features random forest (RF). Compared with only original band model input, RMSE reduced by 0.08 m. The results show that inverted has an 0.45 m compared handheld RTK at mudflat from UAV 0.20 analysis over four-year period, trend towards sedimentation closer land becomes more pronounced.