Objectives  Sunlight pollution caused by glass curtain walls of buildings in urban environment has brought great threats to traffic safety, fire control and residents?? health. At present, most researches on architecture sunlight pollution traditionally focus a local small range at specific construction based on simula-tion or measurement methods. We keep the way of three-dimensional (3D) computation, dissimilarly inte-grates geographic resources to expand the perception range of city sunlight pollution brought by building glasses.

  Methods  The calculation technique of sunlight pollution builds 3D spatial relationship among road, building facade and sunlight in real urban environment, using deep-learning image segmentation to extract glass curtain walls. Then, combined with glare evaluation model, we develop an innovative sunlight pollu-tion model from perspective of drivers to evaluate the suitability for people.

  Results  Taking Shenzhen Fu-tian Center as an example, the multi-level analyses of sunlight pollution are implemented by means of calcu-lation under real time and real location, and path selection based on evaluation. Through these instances, the time sequence regularity, the spatial distribution characteristics and the path estimation of architecture sunlight pollution are revealed.

  Conclusions  The analyses retain space computation mode with innova-tions, fully utilize street view resources, and provide people the macroscopic concept, which indicates rela-tive strength of building sunlight pollution. Additionally, the method ensures instantaneity of urban sunlight pollution acquisition and computation, to a certain extent covers shortages of current researches, and is fur-ther capable of reference evaluation in environment protection, accurate warnings and predictions of glare intensity for pedestrians and drivers.