第七届青年地学论坛

Enriching the Asian perspectives in the research of rapid identification of urban poverty and its implications for housing inequality, this paper contributes empirical evidence about the utility of image features derived from high-resolution satellite imagery and machine learning approaches for identifying urban poverty in China at a community level. In the case of Jiangxia and Huangpi District of Wuhan, image features including perimeter, line segment detector (LSD), Hough transform, gray level cooccurrence matrix (GLCM), histogram of oriented gradients (HoG), and local binary patterns (LBP) are calculated, and 4 machine learning approaches and 25 variables are used to identify urban poverty and relatively important variables. Results show that image features and machine learning approaches can be used to identify urban poverty with the best model performance of R² of Jiangxia and Huangpi of 0.5341 and 0.5324, respectively, although some differences exist among the approaches and study areas. The importance of each variable differs among the different approaches and study areas; however, the relatively important variables are similar. In particular, 4 variables achieved relatively good prediction results for all models, and presented obvious differences in varying community with different poverty level. Housing inequality within low-income neighborhoods, which as a response to gaps in wealth, income, and housing affordability among social groups, is an important manifestation of urban poverty. These findings are useful for policymakers to rapidly identify urban poverty and have potential applications for revealing housing inequality and proofing the rationality of urban planning for building a sustainable society.
 

urban poverty; high-resolution satellite imagery; image features; machine learning approaches; China