Research On Fire-point Positioning Algorithm In The Remote Video Surveillance System Of Forest Ifres

Forest-fire has become a worldwide disaster, which often breaks out suddenly and spreads fast andif not put out without delay, it will cause great damage. So the fire’s early detection and accuratelocation are of great significance for forest-fire prevention departments to put out the fire quickly. Now,there is a mature and effective forest fire prevention technology combining of video monitoring and GIS,which can not only get the real-time pictures of the fire place intuitively, but also access to the terrainfeatures, the forest resource condition and other information around the fire-point. However, fire-pointpositioning based on this technology still lack mature and effective algorithm because of few researcheson it, but forest fire prevention departments have great demand for it.A fire point location algorithm based on forest fires video monitoring system was researched,designed and implemented in this pepper. This algorithm introduced a reference sight window cleverlyto solve the coordinate of the fire-point by fully use of the camera parameters and the geometricrelationships established from the pinhole imaging model of the camera. And the main study contentand results of this thesis are as follows:Firstly, starting from the geometric relationships established by the pinhole imaging model of thecamera, it analysis the relationships among the geographical location of the fire-point, the window, thefire’s image coordinate and the view point for the preliminary design of the algorithm throughintroducing a reference sight window and extracting the relevant parameters of the camera.Secondly, by taking advantage of the camera parameters and geometry knowledge, the solvingmodels of the coordinates of the center point, the boundary points and the fire’s corresponding point inthe reference window and the fire-point sight line model are proposed to get the fire-point’s coordinates.Thirdly, the algorithm of the fire-point location was implemented by Arc GIS Engine develop kitand C#in Visual Studio.NET development environment. Then designed experiments were carried out inJincheng Town, Lin’an City, Zhejiang Province to prove the practicality of the positioning algorithmand the error was analysised by the end. The results showed that the algorithm works well as long as itcan capture the fire images whether the camera is rotating or not in the experimental conditions of thisarticle. And the positioning accuracy is within60meters in X and Y directions, which can satisfy therequirements of the fire location in forest fires protection.The innovation of this thesis performance in the following aspects: Firstly, a fire-point positioning algorithm based on DEM using one single camera is put forwards.Secondly, it’s clever to contact the fire image’s two-dimensional coordinates with the referencewindow’s three-dimensional coordinates to solve the three-dimensional coordinates of fire-point.