The Identification Of Jet Path Of Fire-fighting Cannon Based On Computer Vision

Monitor automatic fire extinguishing system is an important means in today’s fire-fighting, itcan more effectively solve the problem of fire fighting in large space. But because of thecharacteristics of the existing fire water cannon, which is an open loop control of automaticpositioning way, only to the point of orientation, and to monitor the jet trajectory predictionpositioning methods, more vulnerable to changes in the external environment, and resulting in aloss of position precision. In order to realize the monitor in identification and location of jet, thispaper base on the perspective of computer vision, and use image processing technology to identifyfire monitor jet path.Vector search method is proposed in this paper, according the starting point and the initialsearch direction, it can seek the whole jet trajectory step by step. Through vector method, it canconvert the recognition of the jet trajectory, shrink to setting the starting location and the initialsearch direction, greatly convenient the identification process, but also convenient for theintegration and application of the algorithm. In the process of using, it combining the backgroundsubtraction and division, constitute the multiple path search method, combined with particleswarm optimization (pso), constitutes the iterative search algorithm.In the combination with background subtraction division composite more path search method,following the following ways: through the establishment and update of background model, to dothe real-time background deduction, after threshold segmentation and filtering processing,extracting the the target area, which conform to the jet characteristics. Based on the information oftarget area, set up multiple points and the corresponding initial search direction, and then to do themultiple path search, finally by sifting through optimization, get the real jet path. The experimentsprove that in camera fixed condition, this method has strong ability of recognition to the jet, canresist the interference of camera shaking caused by external environment, and can recover fastIn combined with particle swarm optimization constitute the iterative optimization algorithm,setting up fitness value calculation method, according to the movement characteristic of point todo iterative motion to search for the jet path. Preliminary validation, this method has certainfeasibility and the search results have nothing to do with whether background changes, and hasthe strong ability to adapt camera shake cases. Jet recognition algorithm of this paper, is programmed and designed on Visual Studio2010platform, and have done experimental verification for multiple path search algorithm and particleswarm optimization (pso) algorithm respectively, and design software base on multiple pathsearch method around stationary camera situation. The software is designed the user interface byusing WPF framework, and called Emgu CV libraries to assit the written of algorithm, and use themultithreaded way to coordinate image acquisition, processing and interface Interaction. It havegood stable and reliable performance.