Remote Sensing Inversion of Full-Profile Topography Data for Coastal Wetlands Using Synergistic Multi-Platform Sensors from Space, Air, and Ground.

This study proposes a "zonal inversion-fusion mosaicking" technical framework to address the challenge of acquiring continuous full-profile topography data in coastal wetland intertidal zones. The framework integrates and synergistically analyzes data from multi-platform sensors, including satellite, unmanned aerial vehicle (UAV), and ground-based instruments. Applied to the Min River Estuary wetland, this framework employs zone-specific optimization strategies: in the inundated zone, the topography was inverted using Landsat-9 OLI imagery and a Random Forest algorithm (R2 = 0.79, RMSE = 2.08 m); in the bare flat zone, a linear model was developed based on Sentinel-2 time-series imagery using the inundation frequency method, and it achieved an accuracy of R2 = 0.86 and RMSE = 0.34 m; and in the vegetated zone, high-precision topography was derived from UAV oblique photography with Kriging interpolation (RMSE = 0.10 m). The key innovation is the successful generation of a seamless full-profile digital elevation model (DEM) with an overall RMSE of 0.54 m through benchmark unification and precision-weighted fusion algorithms from the sensor data fusion perspective. This study demonstrates that the synergistic sensors framework effectively overcomes the limitations of single-sensor observations, providing a reliable and generalizable integrated solution for the full-profile topographic monitoring of tidal flats, which offers crucial support for coastal wetland research and management.