基于子块匹配法与改进FRFCM的气体泄漏红外图像增强方法  

Infrared image enhancement method of gas leakage based on sub-block matching method and improved FRFCM

作　　者：沈昊旸 于海洋[2] 徐晴晴 董绍华[2,3] 王明月 陈俪赟 SHEN Haoyang;YU Haiyang;XU Qingqing;DONG Shaohua;WANG Mingyue;CHEN Liyun(College of Carbon Neutral Energy,China University of Petroleum(Beijing),Beijing 102249,China;College of Safety and Ocean Engineering,China University of Petroleum(Beijing),Beijing 102249,China;Pipeline Technology and Safety Research Center,China University of Petroleum(Beijing),Beijing 102249,China;Applied Technology Research Institute,Northeast Petroleum University,Daqing Heilongjiang 163318,China;Guangzhou Keyi Optoelectronics Technology Co.,Ltd.,Guangzhou Guangdong 510720,China)

机构地区：[1]中国石油大学(北京)碳中和示范性能源学院,北京102249 [2]中国石油大学(北京)安全与海洋工程学院,北京102249 [3]中国石油大学(北京)管道技术与安全研究中心,北京102249 [4]东北石油大学应用技术研究院,黑龙江大庆163318 [5]广州科易光电技术有限公司,广东广州510720

出　　处：《中国安全生产科学技术》2025年第2期130-138,共9页Journal of Safety Science and Technology

基　　金：中国石油科技创新基金研究项目(2021DQ02-0801)。

摘　　要：为应对红外气体泄漏成像过程中因设备转动等因素导致的图像不稳定及泄漏气体检测效果不佳的问题,提出1种结合图像帧子块匹配法和改进快速鲁棒模糊C均值算法(fast and robust fuzzy c-means,FRFCM)的红外图像细节增强方法。该方法利用图像帧子块匹配法配准图像帧,同时引入背景建模和差分方法从背景中分离动态气体目标,并在FRFCM基础上增加自适应调整模糊因子以优化图像帧的羽流强化特征效果。研究结果表明:该方法能够有效去除冗余信息,使图像帧匹配误差降低约75%,对比度增强值提高4.7%,羽流分割的平均交并比达到0.68,在保持较高分割准确度的同时显著提升检测速度,适用于油气田、集输站及氢气站等气体安全检测系统。研究结果可为气体泄漏监测技术的优化与应用提供参考。To address the issues of image instability and poor effect of gas leakage detection caused by the equipment rotation during the process of infrared gas imaging,an infrared image detail enhancement method combining the sub-block matching method for image frame and an improved Fast and Robust Fuzzy C-Means(FRFCM)algorithm was proposed.The sub-block matching method for image frame was used for image frame registration,and the background modeling and differencing method were used to separate the dynamic gas targets from the background.Additionally,a self-adaptive adjustment fuzzy factor was incorporated into the FRFCM algorithm to optimize the plume enhancement characteristics effect of image frame.The results show that this method can effectively remove the redundant information,reducing the matching error of image frame by approximately 75%,increasing the contrast enhancement by 4.7%,and achieving a mean intersection over union(IoU)of 0.68 for plume segmentation.The proposed method not only maintains high segmentation accuracy,but also significantly improves the detection speed,making it suitable for gas safety monitoring systems in the facilities such as oil fields,gathering stations,and hydrogen stations.The research results can provide reference for the optimization and application of gas leakage monitoring technology.

关 键 词：羽流增强 红外气体泄漏 FRFCM 子块匹配法 背景差分 

分 类 号：X913[环境科学与工程—安全科学]