Research On Acquisition And Retrieval Method Of The Locatable Video

Video,as a kind of public media,with the continuous impact of the image,sound,text and other signals,it brings people a unique visual,auditory enjoyment,which has always been the hotspot of research and development in various fields.In the field of GIS,the video shows the objective world from the side,from the height to the side,the abstract to the image,the specialization to the popularization,breaking the traditional GIS geographic expression,spatial cognition,popularization process and other multiple bottlenecks.However,the video data do not have the real spatial information,so the rich information contained in the video can't be fully utilized,and it is difficult to meet the demand of video and GIS integration development.In this paper,a new locatable video data model which includes video information,audio information,location,lens orientation,shooting posture and shooting parameters.Through the study of spatial information and video synchronization model,coupling technique,and then based on block DCT digital watermarking algorithm to solve the video information and spatial information fusion.On this basis,through the popularity of the Android platform customized locatable video capture system prototype,and verify the feasibility of the model by testing the field data acquisition.For the locatable video data,from the video frame and the fragment,we separately designed the spatial expression model of video position,video track and video viewshed.And with it,we used GeoSOT grid to organization and index the locatable video space object,and construct a one-to-one mapping relationship between "video image-spatial object-grid coding",and establish efficient spatial and temporal indexing mechanism for video frames and video clips,which achieve video-based spatial retrieval.And compared and analyzed retrieval efficiency from different index ways.The experimental results show that the efficiency of indexing based on GeoSOT coding is the highest,which is 20?30 times higher than that of latitude and longitude index,which is 6?8 times of the efficiency of multi-level grid spatial indexing,and the retrieval efficiency increases with the search range.Finally,developed the prototype system for locatable video capture and retrieval based on Android and Web.