Evaluating the Combined Edge Segment Texture (CEST) Method for Damage Detection in Al Fashir, Sudan

This study evaluates the applicability of the Combined Edge Segment Texture (CEST) method for detecting conflict-induced building damage in Al Fashir, Sudan, using moderate-resolution PlanetScope imagery. The research adapts a methodology originally designed for very high-resolution (VHR) data, implementing it through a hybrid workflow that combines Google Earth Engine and Python. The method integrates three analytical branches: edge detection, texture analysis, and segmentation which are then combined using a rule-based decision tree. The results reveal a fundamental incompatibility between the CEST method and the 3–5m resolution of the imagery, leading to a final F1-Score of 39/37%. The analysis found that the mixed-pixel effect and feature blending compromised each analytical branch, resulting in a low recall and a high rate of false positives. While the method lacked the precision for detailed, building-level assessments, it was sufficient at identifying major damage hotspots using visual approach. The findings suggest that the adapted CEST method serves as a tool for rapid, broad-scale damage mapping in a humanitarian context, offering a balance between the affordability of PlanetScope imagery and the need for timely information.