Monitoring Urban Areas by Means of Coherence Levels

Urban areas studies by means of C-band satellite SAR imagery has not provided relevant information while the use of InSAR has shown that it is possible to detect urban settlements with rather good accuracy. In this study we have analysed the potential benefits connected with the combination of four SAR and InSAR signatures (short and long term coherence, backscatter and intensity texture) for urban oriented applications. Two test sites have been considered: Gothenburg, Sweden, and Naples, Italy. A simple algorithm based on thresholds for each parameter used, was able to extract urban land-cover and to provide information on the building density. A visual comparison with optical imagery gave the impression that the extraction was rather accurate. The problem of urban classification was also analysed and results have been promising since different structural categories are characterised by specific values of the four parameters. We may conclude that urban monitoring is possible when SAR and InSAR quantities are combined; nevertheless the accuracy of the method should be defined and more areas should be investigated.