A spatiotemporal segmentation-based framework for mining the spatial and temporal laws associated to urban fire hazards using data obtained through urban grid management is proposed to address the urgent need for preventive measures in urban fire management and prevention. The framework processes spatial and temporal distribution patterns successively. Firstly, spatial distribution patterns of urban fire hazards under different spatial grid resolutions are explored to obtain collections of samples with similar spatially distributed patterns. Subsequently, coexistence patterns and lagging effects of the above-mentioned set of hazards are analyzed to delve into their possible spatio-temporal distribution patterns. Empirical research is conducted with nine typical fire hazard types in Guangming district, Shenzhen, as a case study. The research shows that there are four distinct groups of similarity combinations regarding the spatiotemporal distribution of fire hazards across various spatiotemporal scales. Further spatiotemporal similarity mining indicates the coexistence relationship among hazards, with some hazards exhibiting lagging impact relationships. It can be concluded that the proposed framework is feasible for analyzing and mining spatiotemporal distribution similarity patterns of urban fire hazards. The resulting analytical conclusions of urban fire hazards coexistence, spatiotemporal similarity, and lag effects may provide substantial support for urban fire prevention tasks as well as for other similar studies.

urban fire hazards, spatiotemporal pattern analysis, hazard coexistence and lagging, urban grid management, hazard prevention

针对城市网格化管理形成的城市火灾隐患数据, 基于火灾管理防范措施前置化的迫切需求, 提出一套城市火灾隐患的时空规律挖掘框架。此框架采用时空分治理念, 首先探究不同时空尺度下火灾隐患的空间分布模式, 获得空间分布相似样本集合; 然后基于存在时空分异的火灾隐患数据, 进一步探索时空尺度下的隐患共生模式和滞后影响。以深圳市光明区9类典型火灾隐患为实例开展实证研究, 发现深圳市光明区火灾隐患存在4组时空分布相似组合, 其模式在不同时空尺度下均得到验证; 进一步的时空分布相似性挖掘表明, 隐患组合间存在共生关系, 部分隐患组合间存在滞后影响关系。研究结果表明该框架适用于城市火灾隐患时空分布相似模式的分析和挖掘, 可以输出类间共生、时空相似和滞后性等分析结论, 为后续机理研究和火灾防治等工作提供技术支撑。