Design Of Fire Fighting System For Oil Depot Based On Video Surveillance Platform

With the expansion of the construction scale of oil depots,the operation safety of oil depots has become more and more important,and the fire safety of oil depots is even more important.How to carry out fire detection in time and effectively has become a major problem faced by various large oil depots.Using the existing video monitoring platform to carry out fire signal detection has become a research hotspot of relevant research institutions.Compared with the traditional fire detection equipment,the fire detection technology based on video image analysis has simple installation and large coverage.Using the original monitoring system,there is no need to install additional detection equipment,fire early warning can be realized through real-time analysis of on-site video,the detection accuracy is higher,and the problem of missing reports caused by small installation coverage rate of traditional fire detection technology can be avoided.In addition,the characteristics of convenient fire recording,high accuracy and strong real-time performance also make the fire detection technology based on video image analysis extremely suitable for large-scale places such as oil depots,which are open and difficult to install traditional fire detection equipment in large areas.Video fire detection technology processes video images and recognizes targets according to fire characteristics to complete fire detection.Due to the continuous innovation of computer vision,in-depth learning and other technologies,efficient fire detection technologies emerge one after another and have been widely used in the field of fire safety.Based on the actual problems existing in Wangjiagou oil depot,this paper proposes a fire detection technology based on video monitoring platform,and designs the oil depot fire protection system based on this framework.The system firstly extracts the suspected target area,Then the way of target detection is carried out,Firstly,Gaussian mixture model is used to extract the foreground of moving target,The HSI and RGB color models are combined to continuously track the target to obtain the suspected fire area.In the fire detection part,based on Alex Net depth classification network structure and combined with image processing related knowledge,the model training is carried out to finally obtain the identification model that can identify fire signals such as flame and smoke.At last,through design experiment,The fire protection system designed in this paper is completely suitable for the fire protection work of the Wangjiagou oil depot of western pipeline.It can accurately identify the complex environment,eliminate the interference of external factors,and obtain accurate identification results.It also has strong practicability for other similar fire detection in large space and complex environment.