Article Preview
TopIntroduction
In recent years, with the rapid development of public security big data applications, video surveillance network has become a powerful tool for security management and crime investigation (Guo et al., 2017; Kim & Park, 2021; Nazaré & Schwartz, 2016). The video surveillance network established by the public security department is characterized by large scale of video front-end equipment, wide geographical coverage, diverse system application software, rich bearing business functions and strict monitoring requirements, which helps to improve the efficiency of public security and assist in case investigation (Ma et al., 2018; Thiyagarajan et al., 2018). However, video surveillance systems are exposed to complex network security risks, such distributed networks are easy targets for criminals, and in the event of a major network security incident will result in the unavailability of video surveillance resources, which in turn may threaten the security of the state, society, enterprises and individuals (Ma et al., 2018; Thiyagarajan et al., 2018). Industry and academia have invested heavily in the development of prevention tools and methods to reduce potential losses by moving the prevention barrier forward through monitoring and early warning.
As an important front-end device of video surveillance network, video surveillance cameras are characterized by a large number and variety. If a video surveillance camera is illegally replaced, it can easily develop into a security risk for the whole video network. However, as part of the video dedicated network cameras are leased, purchased, installed and replaced by operators, management departments are often unable to clarify all front-end devices, while the lack of overall security management mechanism for front-end devices, without an efficient front-end device identity authenticity detection method. Therefore, even if video surveillance cameras are illegally replaced, it is often difficult for security management to detect it.
Therefore, in this paper, we propose a video surveillance camera identity identification method that incorporates multi-dimensional static and dynamic identification features. As shown in figure 1, firstly, static identifiers of video surveillance cameras are collected and extracted, including explicit and invisible identifiers. Then a series of dynamic traffic feature sets are extracted starting from the video surveillance camera network traffic perspective. The multidimensional identifiers are fused and the amount of self-information for each identifier is analyzed. Finally, the video surveillance camera identification index is determined according to the contribution amount of the identifier in the identification system. The experimental results prove that the method in this paper can accurately and efficiently perform video surveillance camera identification, which has important theoretical and practical significance.
Figure 1. Video surveillance network workflow
TopSummary
Video surveillance cameras belong to a special IoT device used for image acquisition. IoT device identification and marking technology refers to the identification of device information (type, brand, series, model, etc.) of the target device by means of network sniffing, data mining, etc., and the access in the subsequent process to be authenticated. In the IoT device side access, the security policy based on the physical characteristics of the device avoids the complex and inefficient computation of traditional security policies, and physical information features are generally difficult to forge, and numerous researchers study defense strategies for the perception layer (Islam et al., 2017; Xie et al., 2017). The low computing power of device hardware in IoT systems, energy consumption limits complex sensing networks bring unique security threats and defense requirements for IoT end devices, such as key verification algorithms and cross-domain authentication under low computation (Yang et al., 2021).