Research And Implement Of Security Protection Key Technology In Video Surveillance System

With the development of security video surveillance (SVS) systems, police departments, as well as corresponding administrative offices, are paying more and more attentions to the designs and plans of SVS systems. In considerations of its large scale deployment and lacks of a systematic cyber defense architecture, all deployed SVS systems are facing increasing, persistent, and various cyber security threats.This thesis targets the establishment of a secure security video surveillance system through schemes design on two aspects, including the network security and the video device security. Regarding the network security, in this thesis, we propose a brand new H.264-based key distribution and update algorithm, which is built on our own designed device identity authentication framework and thoroughly changes the common pre-set key schemes widely used in the current SVS systems towards a practical solution enabling device-level mutual authentication and video data encryption adopted in SVS systems. In terms of the device itself security, a device anomaly detection scheme is introduced, which can monitor and alarm the abnormal running states of the video surveillance equipment, such as IP cameras, thereby enhancing the cyber security defense strength. The main contribution of this thesis is summarized as follows.1. In the perspective of network security, a H.264 based key management scheme is designed and implemented in the real-setting SVS system structured under national standard GB/T 28181. We leverage the encoding format, as well as the multimedia data transmission protocol, to distribute and update keys used in the practical secure SVS system, which also provides a practice-facing key management architecture for the video encryption.2. With respect to the device security, we propose a new scheme based on Gaussian background modelling and image feature points matching to detect image anomaly and physical jamming. By using our scheme, IP cameras are able to obtain scene graph automatically and match the historical scene with the current image captured by the camera, thereby being able to detect the physically moving and damages on the IP camera.In this thesis, we investigate the cyber security requirements of the SVS system, design and implement cyber defense techniques aiming at the network and device security. The schemes proposed in this thesis have great values to enhance the cyber security defense strength of the SVS system and are helpful for a comprehensive cyber security architecture of the SVS system.