A Novel Mountain Shadow Removal Method Based on an Inverted Exponential Function Model for Flood Disaster Monitoring

Abstract

Global warming and intensified human activities increase flood disasters, causing annual casualties and economic losses. Mountain shadows are a major source of interference in floodwater extraction from SAR imagery, severely impacting the accuracy of water body detection. This study proposes an innovative approach based on the Inverted Exponential Shadow Removal Model (IESRM). This model can adaptively and dynamically adjust the slope threshold according to the terrain characteristics. It is easy to use, eliminating the need for manual parameter setting. The experimental results demonstrate the following: (1) Water body detection tests across diverse terrains (mountains, plains, and foothill plains) show robust results even in complex foothill regions, with an overall accuracy of 94.51% and a Kappa coefficient of 0.86. (2) A comparative analysis with the shadow formation mechanism method and the HAND (Height Above Nearest Drainage) method revealed that the inverted exponential function model achieved the highest accuracy, with an overall accuracy of 96.46% and a Kappa coefficient of 0.89. The IESRM provides an innovative solution for removing mountain shadows, enhancing SAR imagery-based flood monitoring in complex terrains. It offers timely and accurate data support for flood disaster management agencies.