ANNOUNCEMENTS
With the increasing population growth, urban extent tends to grow outward and vertically upward, high-rise buildings provide an ideal solution to accommodate a large number of people in less area. However, from the perspective of modelers working in the domain of catastrophic response, disaster risk, climate study and wind, it becomes crucial to tap the dynamics of high-rise buildings development over space and time. Remote sensing data, particularly optical and microwave imagery, provides an immense opportunity to evaluate changing urban landscape. Very limited study have been done to detect high-rise buildings using medium resolution satellite image. Shadow and characteristic backscattering value of tall buildings using optical and microwave imagery respectively, have been used as an indicator for high-rise buildings by the existing studies. We used Sentinel-2 and Sentinel-1 dataset in the study. First, we assessed the performance of the existing indices and bands to detect high-rise buildings and shortlisted Combined Shadow Index, (CSI), Shadow Enhancement Index (SEI) and Automatic Water Extraction Index for Shadow (AWEI sh), VV and VH band as the best available indices based on literature. Since none of the final selected bands were able to accurately detect high-rise buildings individually, we formulated a model to detect the high-rise buildings using a combination of microwave and optical imagery. Although the model is tested on several locations to see its global applicability, the accuracy is documented only for the selected study areas. With an accuracy of 88 % and 89 % for Yokohama, Tokyo, and Manila, Philippines respectively, the model not only outperforms the existing methods to identify the spatial location of high-rise buildings but also gives information about their distribution. The output image obtained from the model has values ranging from 0 to 1, where a value close to 1 represents concentrated high-rise buildings cluster and a value near 0 represents sparsely located high-rise buildings. The output obtained from this study is a valuable input in other models related to the urban environment, catastrophic modeling and climate studies.
Keywords: Remote Sensing, Urbanization, Automation, Google Earth Engine, High-Rise Buildings, Disaster Response
