Research On Theory And Information Generation Technique Of Soccer Video Based Psychological War

The rapid development and extensive application of information technology open up a new and wide space for psychological war. At present, psychological war is coming into many aspects of daily life apart from wartime and battlefield, it makes the"multitarget"and"concealment"become two typical characteristics of modern psychological war, however, traditional psychological operation is hard to meet the requirement of modern psychological war because of its inherent limitations, therefore, in this thesis, we explore newfashioned psychological operation that possesses the two typical characteristics.Firstly, the theory of Soccer Video based Psychological War (SVPW) is proposed, after that, the technique for SVPW attack is also analyzed. Secondly, based on the existed technology in soccer video analysis, processing and transmission, the key technologies of information generation in SVPW are investigated. Finally, the above theories and methods are validated by designing and implementing an Information Generation Platform of Psychological Operation (IGPPO). The main contributions of this thesis are as follows,1. We propose the theory of SVPW. At first, modern psychological war and SVPW are compared to indicate the applicability of soccer video. Then, the definition and types of SVPW are discussed, after that, the targets, characteristics, flow and methods of SVPW attack are elaborated.2. In order to accelerate the generation of Psychological War Information(PWI), the flow of soccer video acquiring & sending based on live satellite broadcast, live streaming media in P2P network and video on-demand are proposed. Furthermore, we also design the flow of PWI generation in soccer video and clarify the key technologies within it.3. We present the methods of real-time and automatic position detection, which includes goalmouth localization, billboard detection & recognition, playfield detection and static region detection, for PWI superimposition in soccer video. Different from former rule based methods, machine learning algorithms are applied to discover the latent relationships between low-level features and high-level semantic object, and the precision of position detection can be improved to a certain extent.4. We suggest the methods for PWI superimposition in soccer video --"in-video placement". For the implementation of"in-video placement", the playfield model needs to be predefined, and camera parameters of soccer video need to be determined and tracked in real time between successive frames by matching feature points, after that, geometrical transform and postprocessing are applied to PWI referring to the acquired camera parameters. In this way, the superimposed PWI is vivid and concealed in soccer video.5. We design and implement an IGPPO, which comprises two sub-platforms, i.e. position detection sub-platform and PWI superimposition sub-platform. Subsequently, the application of IGPPO is validated from two aspects, one is"run time"and"concealment"validation in PWI generation, the other is"effectiveness"validation of IGPPO.In conclusion, this thesis provides the theory and attack technique of SVPW, and an in-depth investigation into the PWI generation technique in soccer video. This research is not only the foundation of SVPW, theoretically and technically, but also a positive influence on the development of modern psychological war in information age.