A new urban change detection method based on the local G and local spatial heteroscedasticity statistics

Accurate detection of urban changes is critical for guiding city planning that leads to smart and sustainable development. Few existing methods are able to detect urban changes in a timely manner while preserving great details of spatial heterogeneity. In this article, we propose a new method that combines two spatial statistics, namely the local G and local spatial heteroscedasticity. By jointly analyzing the results of both statistics, we design the Urban Development Index (UDI) to assess the types of urban changes that each spatial unit has been experiencing. Based on the sequences of UDI, we can understand the regional patterns of urban changes by linking to urban morphology. We conduct experiments with both a synthetic dataset and a Rwanda population dataset. The results demonstrate that our method can not only identify completed and ongoing phenomena of urban transition, but also unveil the heterogeneous nature of growth and/or shrinkage inside a city.