Research On Wireless Measurement&Control System And Fire Source Recognition Method Of Forest Fire Truck

Forest fires cause huge loss,which often requires large manpower and materials to be extinguished.Firefighters are extremely exhausted after fire extinguishment,which lets the firefighters hard to perceive forest residual fire or re-ignition.It will lead to greater losses if residual fire is not found and extinguished in time.Therefore,a forest fire engine with wireless measurement and control functions is needed to minimize the loss and effectively avoid casualties of fire officers and soldiers.Two problems as follows needed to solve by the expected forest fire engine:1.Research and development anti-interference remote control operating system to keep firefighters far away from fire;2.Research the technologies including automatic detection&identification of fire source and its environment,target classification and distance identification.The main achievement and conclusions are as follows:(1)Development of wireless measurement and control system for forest fire engineBased on the wireless communication,inertial measurement and motor servo control technology,an wireless measuring and controlling system of the forest fire truck was developed on a LH400CUV type fire-fighting platform,which included wireless communication module,inertial measurement system,laser measuring system,video acquisition and transmission module,speed control module,direction control module and so on.The experiment results showed that the operator can realize the remote control of the vehicle speed,travel direction and fire extinguishing gun in the range of 2000m,and can execute real-time image acquisition of fire scene.(2)Acquisition and pretreatment of laser ranging data in forest fire environmentIn this paper,the data acquisition and pretreatment of laser ranging were carried out,which included experiment design of laser ranging and pretreatment of laser point cloud data.This paper mainly introduced the data acquisition process of on-vehicle SICK laser rangefinder,the method of acquiring point cloud data,the correction and process of point cloud data.The most important problem was to solve the noise reduction in laser signal transmission process,this paper proposed a method based on wavelet transform to ensure time interval accuracy of laser echo signal,the noise of the dynamic pulse laser was processed,the advantages and disadvantages of the soft and hard threshold denoising method were compared,and a new threshold function was proposed.A novel method of wavelet optimal decomposition layer was proposed to improve the accuracy of laser measurement.At the same time,the method was applied to process noise-containing image signal in forest area.(3)Classification and identification method of fire ignition source and its environmental targetFrom the target's point cloud fusion data,the feature of target object was extracted,and a kind of forest fire target classification method was developed,on which based multi-class Gaussian kernel fuzzy support vector machine,the main idea was:in order to solve the semantic uncertainty of the target object(independent object),the visible light data of CCD industrial camera segmentation was used.Based on the linear support vector machine classifier,the shape,color,reflection intensity and three dimension spatial characteristics of fire field in forest environment were improved.In support vector machine,Gaussian kernel function was introduced to improve the accuracy of classification.The results showed that the target classification method of Monocular camera was optimized by the parameters,the comprehensive average correct recognition rate could reach up to 96%,and performed high accuracy of identifying the source contour information,which can meet the requirements of forest fire truck to classify and identify the target object.(4)Automatic detection of fire distance based on high precision infrared binocular imageThe distance identification of fire ignition source based on infrared binocular camera was studied.According to the focal length,parallel optical axis distance,binocular parallax and camera parameters of binocular infrared camera,a set of fire ignition distance identification method was established based on the theory of binocular ranging and error theoretical analysis.The field experiments demonstrated that using laser ranging method,the error was less than 0.5%,while the infrared binocular camera,the distance recognition error of the ignition source was about 5.5%,the accuracy of laser ranging was higher than the one of infrared,but the cost of infrared distance measurement was lower,and the infrared binocular camera could realize the target and distance recognition at the same time.