The Target Video Search Technology Of Monitoring Network Platform Based On GIS

In recent years with the expansion of urbanization, traffic road network presentsa complex state. In order to maintain the social stability; the application of intelligentsecurity is very essential. To obtain multiple video surveillance on the street, thetraditional method is inefficient and a waste of manpower. Combining with theGeographic Information System (GIS) for the target of multiple video search, we canavoid blind search, speed up the search and save manpower. The main contents are asfollows:First of all, a monitoring network platform is set up, which includes on-sitemonitoring terminals and a remote server. The former sends video information to thelatter through internet，so the operator can view the video and search their interestobjectives; The several domestic GIS are compared and the MapX control is selectedas the secondary development of GIS platform. The video decoding, audio decoding,text messaging and other functionality on server-side are described.Secondly, the network modeling of video surveillance on the street is completedaccording to the method of street network modeling. The network model, expressedby topology tables, shows the information of spatial location. MapX control isadopted to set for the GIS. Since the MapX control does not have topologicalrelationships, topological relations should be established and a suitable topologicalmodel should be selected.Thirdly, a research on target video search technology is made. An adjacencymatrix and an judgment matrix are deduced according to the topology tables of arcsand nodes. The target video location on GIS is determined by matrix and direction ofthe target person’s movement. The smallest angle between the neighbor nodes andtarget person is marked and displayed on GIS.Finally, multiple video search algorithm is made and the experimental results areanalyzed. The results show that the complexity of the street network and the level ofcomputer hardware are the effects of the search’s real-time performance. Theaccuracy of the search are influenced by the subjectivity of the operator, theuncontrollability of perpetrator’s action and the region of video surveillance. Thealgorithm has some practical value in the intelligent traffic management, but in terms of the speed and improvement of topological relationship building, furtherimprovement can be done.