Random‑forest‑based bathymetric inversion code for shallow lakes on the Tibetan Plateau using Sentinel‑2 imagery

It implements the random‑forest‑based bathymetric inversion workflow, including multi‑scale spatial‑statistical‑feature extraction with the kernel‑median method, random‑forest regression prediction, smoothing via median and Frost filtering, and export of bathymetric rasters for plateau lakes.

No remote‑sensing imagery, in‑situ sample datasets or other auxiliary geospatial files are distributed with this code package; users must prepare their own input data.Implementation environment All scripts are written in MATLAB, and all core computations are completed within the MATLAB platform.