Abstract:Comprehensive research on the urban green visible index and the structural characteristics of street trees can provide a reference for improving the urban greening effect and ecological environment quality as well as provide a scientific basis for urban greening planning. A total of 40 streets in 6 central blocks of Anning District of Lanzhou were selected for this study. ArcGIS and Python software were used to obtain Baidu Street View images, ImageJ or Sketchup was used to obtain street tree structural feature data, and the DeepLabv3+ model was used to obtain green view rate data and conduct quantitative analysis. The results showed that the most frequently used street greening in the Anning district was Styphnolobium japonicum. Trees with a height of 5 m~10 m reached 75.5%, trees with a crown width of 4 m~8 m reached 58.3%, and trees with a diameter at breast height greater than 45 cm only 2.3%. Streets with low green vision accounted for 75.3% of the study area, while streets with high green vision accounted for only 7%. The changes in plant growth indicators such as the tree height, crown width and diameter at breast height will affect the leaf canopy and thus affect the green visible index, and all of them are positively correlated with the green visible index, which affects the visual experience and psychological feeling of residents to a certain extent. Based on the conclusion, it is suggested that the streets of Peili with high green vision should be trimmed and maintained regularly, and the streets of Kongjiaya and Anningbao with medium and low green vision should be planted with native tree species to improve the green vision.