A framework for the evaluation of roof greening priority

Under increasingly low urban land resources, greenery on buildings is an innovative approach used to solve urban and architectural environment problems and improve urban green spaces. Therefore, it is necessary to evaluate the priority of roof greening for buildings in cities. And, a comprehensive evaluation in urban areas is required. This study assessed the priority of roof greening from two perspectives: capabilities and requirements. Deep-learning method (YOLO V3) was used on remote sensing images to estimate the capability for detecting candidate buildings for roof greening, while temperatures, rainfall, park or green space data, and road congestion data were used to measure the required indicators. In order to comprehensively evaluate the roof greening priority, an assessment method was constructed using the weighted sum of multiple indicators. Xia’Men Island, a highly urbanized area, was used to validate the proposed method. Based on the priority assessment model developed herein, we measured 956.64 ha that had roof greening potential. Space and time guidelines can provide references for effectively promoting project implementation and improving sustainable urban development. •Deep learning method is used to select candidate buildings for roof greening;•A comprehensive framework was proposed for assessing roof greening priority;•Quantitative and Qualitative analyses of Xia’Men Island were conducted.