Research On Intelligent Fire Monitor Control Based On Machine Vision

As the primary equipment for long-distance fire-fighting,fire monitors hold advantages like large jet flow and flexible transportation position,which is one of the key fire-fighting equipment for fire-fighting and rescue,rapid cooling of fire sites,and curbing secondary disasters.With the development of various emergency rescue equipment in the direction of high mobility and multi-function,and faced with requirements of various fire suppression and fire protection scenarios,fire monitors are also developing into subdivision applications directions.Existing fire monitors still mostly rely on firefighters’ operation and control.Some of the fire monitors used in large-space automatic fire extinguishing systems and fire robots have low reliability and poor functionality,which makes it difficult to meet the application requirements of various fire-fighting scenarios.Therefore,it is necessary to carry out research on intelligent fire-fighting monitors for fire-fighting and rescue.And with the help of increasingly mature computer vision technology,through control of the pitch and yaw angle,the fast,continuous and stable conveying of water jet trajectory to fire position in the process of fire-fighting could be realized,and the efficiency of fire-fighting and ensure safety of firefighters could be improved to the maximum extent.This project is supported by the National Key R&D Program of China “Key Technology Research and Product Development of New Multifunctional Urban Fire Fighting Vehicles”(No.2016YFC0802908),and with fire monitors as the research object.This work combines related theories and technologies such as fluid kinematics simulation,computer vision,image feature extraction,jet trajectory parameter characterization,and intelligent control,including parameterized modeling of jet trajectory,online prediction of trajectory falling position,position compensation for fluid supply pressure fluctuations,and fire identification and tracking fire suppression.The proposed fire extinguishing and rescue solution provides theoretical support and principle verification for the realization of intelligent fire monitor in the intelligent application of fire rescue equipment.The main contents include:(1)Based on the analysis of current electric fire monitor assembly structure,intelligent fire monitor solution based on computer vision was proposed,and the technical route to realize the pitch and yaw angle control of fire monitor with near-field vision and infrared vision respectively was elaborated.By comparing and analyzing the trajectory parameter error of simulation method and experimental measurement,a jet trajectory parameterized modeling method based on trajectory simulation data supplemented by experimental measurement data error compensation was proposed to realize rapid adjustment of the pitch angle at the initial stage of fire-fighting.(2)In order to realize continuous and effective coverage of the fire position of the water jet trajectory falling during fire extinguishing process of the intelligent fire monitor,an online jet trajectory monitoring study based on Near-Field Vision was carried out.According to the relationship between water jet trajectory motion characteristics and its image features,a water jet state characterization method based on the image features of initial trajectory was proposed,and a joint image processing algorithm including near-field image preprocessing,trajectory segmentation,trajectory equation fitting and feature extraction was designed.Finally,based on the water jet trajectory parameterized model developed in this dissertation,the online water jet trajectory falling position prediction of fire monitor during jetting was realized.(3)Focusing on the horizontal aiming during the fire extinguishing process of the intelligent fire water monitor,research on fire identification and tracking fire extinguishing based on infrared vision was carried out.Firstly,the correlation between infrared image temperature and circularity characteristics and fire detection distance was analyzed,and a joint fire recognition method based on temperature and circularity characteristics was developed.Secondly,the intelligent fire monitor yaw control algorithm that simulates firefighter’s aiming and tracking fire-fighting operation mode was established.Finally,experimental simulation of multi-scene fire extinguishing applications requirements including single fire aiming,single fire tracking and multi fire aiming was carried out.Through the adjustment of fire monitor yaw angle,effective fire identification and aiming,fire tracking and multiple fire sequential fire extinguishing was realized.(4)In order to solve the problem of jet trajectory falling position interference caused by fluid supply pressure fluctuations in the application of intelligent fire monitors for fire robots,the study of jet trajectory position compensation based on near-field vision continues to be carried out in this article.Firstly,considering the complex and changeable background of near-field image during mobile fire-fighting process of fire robot,a new near-field image trajectory segmentation method was proposed and the jet trajectory feature extraction process was optimized.Secondly,jet trajectory falling position compensation model was developed by means of support vector regression(SVR)based on pressure fluctuation trajectory characteristics and falling position data.Finally,two conventional pressure fluctuation scenarios were simulated: fluid supply pressure decreases then recovers and rises then recovers,respectively.Through the adjustment of fire monitor pitch angle,the continuous and effective coverage of fire location under the fluctuating liquid supply pressure was realized.Finally,an automatic fire-fighting experiment was carried out and the results show that the machine vision-based fire water monitor control scheme can achieve the continuous and accurate coverage of the fire source by water jet trajectory.There are 80 figures,15 tables and 188 references in this dissertation.