烟雾图像分割下的大型综合体最优化消防疏散路径循环模拟  

Optimal Fire Evacuation Path Cycle Simulation for Large Complexes under Smoke Image Segmentation

作　　者：劳晓镜 Lao Xiaojing(Zhejiang Greenton Architectural Design Co.Ltd.,Hangzhou 310012,China)

机构地区：[1]浙江绿城建筑设计有限公司,杭州310012

出　　处：《绿色建造与智能建筑》2025年第4期124-128,共5页Green Construction and Intelligent Building

摘　　要：在大型综合体设计中,存在消防疏散路径复杂、耗时较长的问题,由此,设计一种烟雾图像分割下的大型综合体最优化消防疏散路径循环模拟方法。基于深度学习,设计一种由语义边界检测网络与语义分割网络构成的EANet模型,实施大型综合体火灾中的烟雾图像分割,根据烟雾分割图像与建筑图像,采用栅格标识法与自适应蚁群算法划分大型综合体的建筑空间平面,构建建筑疏散空间模型。考虑火灾灾害实时扩散因素,改进传统蚁群算法,实现最优化消防疏散路径的循环模拟。仿真测试结果表明,该方法能够实现各处的最优化消防疏散路径循环模拟;在起火点有两处时,仍然可以保持较短的路径循环模拟时间。仅需较短时间即可实现准确的最优化消防疏散路径循环模拟,有着良好的实用性。In the design of large complex,thefire evacuation path is complex and time-consuming.Therefore,a circular simulation method of optimal fire evacuation path of large complex under smoke image segmentation is designed.Based on deep learning,an EANET model composed of semantic boundary detection network and semantic segmentation network is designed to segment the smoke image in the fire of large complex.According to the smoke segmentation image and building image,thebuilding space plane of large complex is divided by grid identification method and adaptive ant colony algorithm,and the building evacuation space model is constructed.Considering the real-time diffusion of fire disaster,the traditional ant colony algorithm is improved to realize the circular simulation of optimal fire evacuation path.The simulation test results show that this method can realize the cyclic simulation of the optimal fire evacuation path everywhere The simulated path can be maintained for a short time at two points.It takes only a short time to realize the accurate optimal circular simulation of fire evacuationpath,whichhas good practicability.

关 键 词：烟雾图像分割 大型综合体 语义边界检测网 最优化消防疏散路径 路径循环模拟 栅格地图 

分 类 号：U458[建筑科学—桥梁与隧道工程]