Detection of urban built-up areas by using Sentinel-1images from different orbits, Case study: Isfahan

In recent decades, built-up urban areas have expanded rapidly as a result of population growth and economic development. In developing countries, this trend is faster. It is essential to Know the trend of rapid land-use changes for urban managers to plan for the future growth of the city while providing appropriate urban services. Satellite imagery is a reliable source in built-up areas extraction. Among the various types of satellite imagery, radar imagery is effective in urban areas extraction because they captured images in all weather conditions and ascending and descending orbits. In this study, the performance of the time series of ascending and descending images of Sentinel 1 in VV and VH bands were evaluated in the extraction of built-up areas. The areas with high slopes were masked using a digital elevation model to reduce the effects of geometric distortions. The threshold of the built-up areas was extracted from the image histogram using the Otsu automatic threshold algorithm. The results were further evaluated by a high-resolution Google Earth image. In both polarimetric bands, the image in descending orbits has higher overall accuracies in comparison to ascending orbits. The overall accuracies in VV and VH were 90% and 87% in the descending orbit and 88% and 84% in ascending orbit, respectively. The findings of this study show that the VV image has higher accuracies in both orbits in comparison to the VH image. The descending image in VV has 90% overall accuracy in urban area extraction in Isfahan city.