| Forest fire prevention and fighting is an important challenge in the field of emergency fire fighting.Forest fire is prone to cause a variety of negative effects on forest ecosystem,including destroying forest resources,causing geological disasters,forming environmental pollution and so on.However,the current forest fire detection and fire extinguishing work mainly rely on manual operation,which is very dependent on fire detection experience of fire personal and response ability of fire extinguishing organization.The degree of automation is low,which seriously threatens personal safety.In order to speed up the fire prevention and reduce the loss of resources and property caused by forest fires,this thesis puts forward the fire prevention and extinguishing scheme of forest water cannon tower,and studies the jet perception and fire extinguishing control method of water cannon under the overhead shooting condition.The main research content of this thesis is as follows.(1)The working conditions of forest water cannon tower in forest environment is analyzed,its structure is designed and the working requirements is put forward.In order to avoid trees blocking,the visual perception module and the water cannon head are installed on high and shot jet overhead.The hardware components of control system is selected,and the construction of the water cannon tower experimental platform is completed.(2)The jet movement of water cannon under overhead shooting condition is analyzed and the relationship between jet drop point and exit angle is studied.The jet trajectory under particle motion state is analyzed,and the difference between rising and falling state of jet is compared.The motion force of jet under the condition of air resistance is analyzed,and the calculation method of air resistance is determined by combining with the principle of jet enrolling,then the prediction model of overhead jet landing point is established.(3)The real-time monitoring method of jet landing point based on YOLOv5 is proposed.The image data set of jet landing point is collected and established,the image sample data is enhanced and the jet landing point recognition model is trained.The grayscale cooccurrence matrix is used to calculate the texture features of the image,and the relationship between the image information content and the jet brightness is analyzed to obtain the optimal brightness interval of the jet image.With the principle of camera exposure adjustment,a method of automatic brightness adjustment of jet image is proposed.(4)Combined with jet perception method,fire fighting control method of the forest water cannon tower is proposed.According to actual needs,fire fighting work can be divided into open-loop fast approaching phase and the close-loop continuous hitting stage.The spatial distance of the target fire point is measured by binocular perception,combined with the prediction model of the overhead jet drop point,the water cannon tower is open-loop controlled to approach target fire point quickly.Combining the real-time monitoring method and particle filter method,the real-time position of jet landing point is tracked.With the reasonable deviation value of pixel setting as the control dead zone,the close-loop control of forest water cannon tower is realized based on the image servo control method.Considering the fire moving,the moving target position is predicted by Kalman filter in combination with jet landing time,then the forest water cannon tower is controlled to track the moving target by jet.(5)The software of fire extinguishing control system of water cannon tower is developed based on the experimental platform and its control system hardware device.The feasibility of automatic adjustment of jet brightness,open-loop control,closed-loop control and tracking-hitting of moving target are verified by experiments.The experimental results prove the effectiveness of the fire fighting control method that proposed in this thesis.There are 66 figures,14 tables and 94 references in this thesis. |
PDF Full Text Request