Development of a stream DTM generation methodology using UAV-based SfM and LiDAR point cloud.

Area of Science:

• Geospatial Science
• Hydrology
• Ecology

Background:

• Accurate Digital Terrain Models (DTMs) are essential for hydrological and ecological modeling in riverine environments.
• Generating precise DTMs in stream areas presents unique challenges due to water bodies and dense riparian vegetation.
• Existing methods require refinement to effectively integrate diverse point cloud data for improved DTM accuracy.

Purpose of the Study:

• To propose and evaluate a methodology for generating accurate DTMs in stream areas using integrated UAV-based Structure from Motion (SfM) and LiDAR data.
• To compare the performance of three different ground filtering algorithms (CSF, PTIN, SMRF) for removing non-ground points in riverine environments.
• To assess the accuracy of the generated DTMs for hydrological and ecological modeling applications.

Main Methods:

• Integration of point clouds acquired via UAV-based LiDAR (Zenmuse L1) and SfM (Phantom 4 multispectral).
• Classification of water and non-water areas using the Normalized Difference Water Index (NDWI).
• Application and comparison of three ground filters: Cloth Simulation Filter (CSF), Progressive TIN (PTIN), and Simple Morphological Filter (SMRF).

Main Results:

• The Simple Morphological Filter (SMRF) demonstrated superior performance, achieving a Mean Absolute Error (MAE) of 0.160 m and Root Mean Square Error (RMSE) of 0.214 m.
• SMRF effectively removed vegetation and preserved riparian terrain features.
• SMRF showed a tendency to underestimate elevations in water areas, indicating a need for further refinement for submerged terrain.

Conclusions:

• The proposed methodology, integrating UAV-LiDAR, SfM, and advanced filtering, significantly enhances DTM accuracy in stream environments.
• SMRF is identified as the most effective filter for DTM generation in this study, balancing vegetation removal and terrain preservation.
• The improved DTMs offer substantial benefits for mesoscale hydrological and ecological modeling in small riverine systems.