| Forest fire prevention has always been the top priority of forest resources protection.How to effectively monitor and accurately locate forest fires is the core of achieving scientific decision-making and rapid extinguishing.At present,the forest fire monitoring and positioning technology has various forms.Among them,video monitoring is the most widely used monitoring method in forest fire monitoring technology.The video monitoring system is mainly based on the GIS system,and generally the parameters such as azimuth and pitch angle are obtained by PTZ to calculate the forest fire location.However,the positioning accuracy of this method is limited by the hardware performance index.The Geo-tagged image is a real image with geo-location information obtained by using an image acquisition device with built-in GPS function.With the help of the geo-information contained in this kind of image,it can be applied to image-based positioning and enhanced expression of 3D scenes.Firstly,this paper transforms the geo-tagged image into a positionable image based on the principle of terrain matching technology,applies this kind of technology to forest fire prevention,and then proposes the forest fire positioning method based on geo-tagged image,and the positioning accuracy is analyzed and evaluated in this method.Then this paper explores the realization process of the three-dimensional visual representation of the fire image based on geotagged image,and the geo-tagged image is projected as a texture onto the three-dimensional surface.Finally,the Jiulongshan Nature Reserve in Beijing is taken as the research area,and the positioning method proposed in this paper is verified and displayed.The main research content is as follows:1.Smoke and fire positioning based on terrain matching: this paper converts the geo-tagged image into a positionable image by using the principle of terrain matching,and proposes a smoke and fire method based on single-viewpoint geo-tagged image.Aiming at the corresponding positioning area on DEM of the pixel selected in positioning area of image,this paper proposes the analysis and evaluation process of the accuracy information of the positioning area.According to the analysis results,the process divides the positioning area into a positionable area and an untargetable area,and provides detailed position information of smoke and fire.2.Multi-viewpoint geotagged image positioning: Based on the smoke and fire positioning method of single-viewpoint geotagged image,this paper explores the realization of the smoke and fire positioning process of multi-viewpoint geotagged images,and classifies the overlay forms of positioning areas gotten by the positioning of two viewpoints respectively,and finally dicusses the method for obtaining positioning accuracy information from the intersection areas gotten by several possible overlay types.3.3D visualization of fire field based on geotagged images: Using the mapping relationship between multi-viewpoint geotagged images and DEM,this paper projects the real images captured from multiple observation angles onto DEM 2D terrain,and then mosaics and fuses these images which have overlap,and finally takes the fused image as a texture to the DEM three-dimensional terrain,in order to achieve a three-dimensional visual representation of the fire field based on the geo-tagged image.This article mainly includes the following innovations:(1)Applying geotagging images to forest fire positioning research,proposes a smoke and fire positioning method based on geotagged images.The positioning accuracy of this method is only related to the image resolution and the matching accuracy between virtual terrain and real image,avoiding that the traditional PTZ positioning is limited by the hardware performance index.(2)Establishing the conversion relationship between multi-view geographic images and DEM,and realize the three-dimensional visual representation of fire based on real images through projection,mosaic and fusion. |
PDF Full Text Request