基于视觉特征匹配技术的火灾疏散路径红外热成像规划方法  

Infrared thermal imaging planning method of fire evacuation path based on visual feature matching technology

作　　者：马雨晴 李丽华[1] MA Yu-qing;LI Li-hua(School of Art and Design,Chongqing Jiaotong University,Chongqing 400044,China)

机构地区：[1]重庆交通大学艺术设计学院,重庆400044

出　　处：《贵阳学院学报（自然科学版）》2023年第2期96-100,105,共6页Journal of Guiyang University：Natural Sciences

基　　金：重庆市教育委员会科学技术研究项目“后疫情时期城市轨道交通站点人性化空间设计研究与应用”(项目编号:KJQN202000739)。

摘　　要：火灾疏散路径规划时,规划过程易受复杂背景、噪声、光照强度等问题的干扰,导致火灾疏散路径规划准确性较差。为此,提出基于视觉特征匹配技术的火灾疏散路径红外热成像规划方法。剔除火灾热红外图像中的噪声,采用自适应邻域算法增强图像的细节特征,利用视觉特征匹配技术获取火源点,构建路径规划模型,采用遗传蚁群算法对模型求解,获得最佳路径,完成火灾疏散路径红外热成像的规划。实验结果表明,所提方法的规划时间为0.52 s,规划的路径长度短、远离着火点,且能够精准避开所有障碍物。During fire evacuation path planning,the planning process is susceptible to interference from complex backgrounds,noise interference,light intensity,and other issues,resulting in poor accuracy of fire evacuation path planning.Therefore,an infrared thermal imaging planning method for fire evacuation path based on visual feature matching technology is proposed.Eliminate the noise in the fire thermal infrared image,use adaptive neighborhood algorithms to enhance the detailed features of the image,use visual feature matching technology to obtain the fire source point,build a path planning model,use genetic ant colony algorithm to solve the model,obtain the best path,and complete the planning of the fire evacuation path infrared thermal imaging.Experimental results show that the proposed method has a planning time of 0.52 seconds,a short planned path length,is far from the ignition point,and can accurately avoid all obstacles.

关 键 词：抑制噪声 局部对比度增强 距离相似度 适应度函数 信息素变化 

分 类 号：TN96[电子电信—信号与信息处理]