基于多传感器融合的隧道智能巡检系统  

Tunnel Intelligent Inspection System Based on Multi-sensor Fusion

作　　者：朱鹏浩 张世义 李军[1] ZHU Peng-hao;ZHANG Shi-yi;LI Jun(School of Mechatronics and Vehicle Engineering,Chongqing Jiaotong University,Chongqing 400074,China;School of Shipping and Naval Architecture,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]重庆交通大学机电与车辆工程学院,重庆400074 [2]重庆交通大学航运与船舶工程学院,重庆400074

出　　处：《科学技术与工程》2023年第2期648-655,共8页Science Technology and Engineering

基　　金：国家自然科学基金(51305472);重庆市研究生联合培养基地(JDLHPYJD2018003)。

摘　　要：针对交通隧道日常维护及异常情况的处理多依赖于人工巡检所存在的效率低、成本高等问题,提出了一种基于多传感器融合的隧道智能巡检系统。首先,将数据采集功能进行模块化处理,以充分获取隧道内的环境信息;其次,将基于粒子群算法优化的前馈神经网络火灾智能监测技术应用到隧道火灾检测模块中;最后,结合帧差法交通事故图像识别技术,使隧道内异常情况的识别与处理效率得到了明显提升。结果表明:隧道智能巡检系统能较好地识别隧道内的情况及灾险,并可以及时进行报警处理,从而有效地避免二次交通事故的发生。同时系统可根据其他类型隧道的实际检测环境进行定制化设计,具有较高的泛化能力。Aiming at the daily maintenance of traffic tunnels and the handling of abnormal conditions,which mostly depend on the low efficiency and high cost of manual inspection,a tunnel intelligent inspection system based on multi-sensor fusion was proposed.First,the data acquisition function was modularized to fully obtain the environmental information inside the tunnel.Secondly,the back propagation neural network fire intelligence monitoring technology based on particle swarm algorithm optimization was applied to the tunnel fire detection module.Finally,the combination of the frame difference method traffic accident image recognition technology enabled the recognition and processing efficiency of abnormal conditions in the tunnel to be significantly improved.The results show that the tunnel intelligent inspection system can better identify the situation in the tunnel and the disaster risk,and can make timely alarm processing,thus effectively avoiding the occurrence of secondary traffic accidents.At the same time,the design can be customized according to the actual inspection environment of other types of tunnels,with high generalization capability.

关 键 词：隧道智能巡检 多传感器融合 多层前馈神经网络 粒子群算法 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]