A methodological workflow for ultra-high resolution DSM generation with LiDAR data in wetlands

In the framework of wetland biodiversity crisis, advanced remote sensing technologies (based on ultra-high-resolution sensors), offer valuable solutions for monitoring aquatic vegetation – a key component of these ecosystems. Effective image orthorectification at centimetric resolution relies on accurate Digital Surface Model (DSM). The limitations of photogrammetry based on GCPs are particularly evident in aquatic environments, where GCP geolocation can be logistically difficult, and tie points matching typically fails over water. Light Detection and Ranging (LiDAR) can overcome some limitations providing dense 3D point clouds with set density. Nevertheless, shallow bottom in shortwave and strong water absorption in near-infrared cause LiDAR data distortions and gaps over water. This study proposes a semi-automatic workflow for LiDAR point cloud classification and DSM reconstruction over sites with terrestrial and aquatic targets, and evaluate its performance on real-world scenarios