Spatio-Temporal Urban Change Mapping With Time-Series SAR Data

Abstract

The strong urbanization impetus of developing countries leads to various urbanization phenomena such as building constructions, reconstructions, and demolitions. It is desirable to monitor and recognize these intraurban changes by utilizing temporal and spatial information in an automatic way. This may be useful, for example, to timely update urban information databases. The aim of this work is, therefore, to automatically extract first, and further recognize, change time series in sequences of SAR data with high-frequency acquisition. Specifically, SAR time-series segmentation and unsupervised classification are combined together to recognize areas with the same urban change pattern, by fully exploiting both the temporal and spatial dimensions. Experimental results in a fast-growing Chinese city show that the proposed approach is effective and able to characterize temporal patterns due to different kinds of intraurban changes.