城市街景环境与社会安全事件的空间关系研究*

为研究城市街景环境与社会安全事件之间的空间关系，提出融合语义分割技术与地理加权回归方法，基于国内典型城市街景环境数据和社会安全事件数据，验证方法的可行性。研究结果表明:模型校正拟合因子R2在0.744～0.841之间，拟合效果较好，街景环境要素指标与社会安全事件存在显著空间相关性;绿视指数和天空开阔指数与社会安全事件在大部分区域呈显著负相关关系，交通信号指数、汽车指数和人群聚集指数与社会安全事件呈显著正相关关系,研究结果可为城市安全规划和社会安全风险防控提供决策支持。

Research on spatial relationship between urban streetscape environment and societal security events

In order to study the spatial relationship between the urban streetscape environment and societal security events,an analysis method coupled semantic segmentation technology with geographically weighted regression was proposed,which was validated by using the streetscape environment data and societal security events data in a typical city of China.The results showed that the correction fitting factor R2 of this model was between 0.744 and 0.841,which indicated a good fitting performance.Furthermore,it was found that there existed a significant spatial correlation between the streetscape environment element indexes and the societal security events.Specifically,the Green View Index and Sky Openness Index presented a significant negative correlation to the societal security events in most regions,while the Traffic Light Index,Car Index and People Gathering Index presented a significant positive correlation to the societal security events.The results can provide decision-making support for the urban security planning and the societal security risk prevention and control.