Camera Network Topological Mapping With Physical Distribution And Network Information

Camera network mapping is an important part of camera network security,which can help managers to collect camera information and manage camera network.Due to the expansion of camera application scenes and the increase in the number of camera,the mapping of camera network becomes more and more complex.At the same time,many cameras are in the closed environment of the intranet,and the existing network surveying and mapping methods cannot meet the demand of the intranet camera surveying and mapping.This paper focuses on the discovery of the network topology of the internal camera.Based on the idea of mutual relation between the physical space distribution and the network space topology,a camera network topology mapping method combining the physical distribution and network information is proposed to solve the problem that the network structure of the internal camera is difficult to detect.On how to determine physical distribution between the camera and use the physical distribution to discover camera network connection relationship problem,this paper based on pedestrian trajectory judging camera physical distribution and the objective fact that the cameras in the same subnet are closed,proposes the method to judge the camera distribution based on the pedestrian transfer probability and the method to speculate the network connection relationship between cameras based on the physical distribution.The method proposed in this paper improves the efficiency of video content analysis by only recording the events that pedestrians get in or out the camera shooting range and data filtering.Compared with other methods for analyzing the physical spatial distribution of cameras,the proposed method can achieve an accuracy rate of 80% in a shorter time.As some cameras in the intranet are located in remote areas,few pedestrians are seen in the video content captured,so it is difficult to obtain its physical distribution information and judge its network space connection status.For the problem that remote camera could not judge the subnet of the nodes by physical distribution,based on thinking that the network information,such as local network latency,packet loss of network information and the hop,has the similarity,this paper put forward a kind of subnet node belonging judgement method based on bayesian inference to survey and map intranet cameras topology.To ensure the accuracy of the calculation results,an evaluation method based on Bloom Filter and Jaccard similarity coefficient is designed.For a real Intranet camera network scene,using the method designed in this paper,the accuracy of the mapping results reached more than 82%.Based on the topological mapping method of internal camera network proposed above,the corresponding system is designed and developed on the self-developed video monitoring platform of safe city.The function verification of the method proposed in this paper reflects the feasibility and effectiveness of the method proposed in this paper.