Harnessing multiple data sources and emerging technologies for comprehensive urban green space evaluation

Abstract

Urban greening is vital for environmental sustainability and human welfare, but assessing its impact can be challenging due to traditional methodologies' limitations. These methods usually focus on singular metrics, giving a constrained understanding of urban green spaces. This study presents an innovative methodology that overcomes traditional research constraints, providing a comprehensive view of urban green spaces. The approach integrates longitudinal and cross-sectional assessments, remote sensing images, and Google Street View data. It leverages geodata processing and semantic segmentation models for key metrics—green coverage rate and green view index—while classifying various landscape types through an image model. Osaka City has been selected as the study's exemplar to integrate topographic variables and validate this novel methodology. The study results showed that Osaka has conditions such as uneven distribution of green spaces and different quality of green spaces. Targeted suggestions such as focusing on areas with low green visibility and improving the quality of existing green spaces need to be developed accordingly based on the results, paving the way for strengthening the city's urban greening initiatives. Subsequent research must tackle limitations like data quality and model accuracy, spotlighting the crucial role of green spaces in sustainable urban development.