Integration of Sentinel-1 and Sentinel-2 data for change detection: A case study in a war conflict area of Mosul city

The city of Mosul located approximately 400 km North of Baghdad was captured by the Islamic State during the period 04th – 06th of June in 2014. The impact of conflict had devastated results on buildings and the infrastructure of the city.Remote sensing technology has proven an extremely useful tool for monitoring changes. In this study, a combination of Synthetic Aperture Radar (SAR) and optical imagery is used to map and monitor changes in the urban environment due to war.In total, twenty of both Sentinel-1 and Sentinel-2 images were collected during the period of 2015 and 2019. The data were pre-processed and integrated to develop a supervised machine learning classification approach using Support Vector Machine (SVM) classifier. Texture analysis such as Grey Level Co-occurrence Matrix (GLCM) was also performed and combined with the other datasets.Spectral indices have been used to improve the classification results such as: Band Ratio Built up Area (BRBA), New Built-up Area (BUI), Dry Bare Soil Index (DBSI), Modified Normalized Difference Water Index (MNDWI), Normalized Difference Vegetation Index (NDVI), and Soil Adjusted Vegetation Index (SAVI) which were derived from the optical data (Sentinel-2) to serve as an input to the classification scheme. The built-up area index was found to be one of the most important features in the classification scheme.show an overall accuracy of approximately 96% was achieved for all the scenes between 2015 and 2019. The change detection results show that the largest devastation occurred between 2017 and 2018 where the urban area was decreased by 427.65 km², where in 2019 the impact of the Mosul city was the lowest with a 296.8 km² urban loss. The developed approach showed an effectiveness in mapping the areas where conflict had the strongest impact.