The integration of urban streetscapes provides the possibility to fully quantify the ecological landscape of urban green spaces: A case study of Xi’an city

Urban green space is the critical composition of urban ecosystem, which provides numerous ecosystem services for urban residents, and has great value. The optimization of urban green landscape aesthetics can improve the health level of the city and enhance the enjoyment of the people. In this research, we calculated two kinds of ecological indices based on deep learning semantic recognition and multispectral remote sensing, and proposed a set of integrated multi-dimensional urban green ecological indicators, which can be used to evaluate the value of urban green ecological landscape aesthetics. These two kinds of ecological indices represent the landscape characteristics of urban green space from the perspective of overlooking and the perspective of looking around respectively. And then the regionalization method based on dynamic constraint aggregation cluster and partition (RECDAP) is used to identify the ecological homogeneous area of urban block. The results indicated that first, there are spatial differences in pedestrian perception of urban vegetation, and high-perception areas are mainly distributed in schools, parks and internal roads in large communities. Second, the green cover based on overlooking view is also mainly distributed in schools, parks, large communities and other places. Finally, the integration of indicators proves that human perception of urban green space landscape aesthetics cannot be ignored. The homogeneous areas provide a new reference for urban managers and designers to create and maintain urban green space. The addition of streetscape provides a possibility for a more objective and comprehensive evaluation of the aesthetic value of urban green space landscape, so that people can better understand the dynamic role of green space in the urbanization and human health. Public managers and urban green designers can use the ecological indices system to optimize and maintain urban vegetation by taking ecologically homogeneous areas as units.