Survey Of The Key Technologies For Security Surveillance Video Perceptual Noise Blind Detection

Security construction in full swing, plays a pivotal role in the country to create asafe city, the process of building a harmonious society. With the rapid development ofvideo surveillance technology, users demand for higher requirements on image quality,function perfection and performance stability. However, on collecting the video source,the encoded video signal during compression and transmission, the actual noiseperceptual interference will result in different levels of video quality degradation,therefore monitoring of real-time detection of video quality and repair becomesimportant subject at this stage to be solved.In response to these needs, the paper maked security surveillance video of theperception of noise as the study, and analyzed the perception of noise in the videotime-domain characteristics. Focusing on how to introduce the concept of fuzzy set ofvideo image processing, the paper overcomed the actual surveillance video detect in thepresence of noise. The main work includes the followings:Firstly, summary reviewed the development of security video surveillance system,summarized the main relevant video image noise detection algorithms and theiradvantages and disadvantages, and the current research status of video noise detectionalgorithms. Secondly, the paper described causes and types of different surveillancevideo noise causes and maked full use of the time-domain-related features of videosequences studied perceptual video noise and its characteristics in the time domain, andmainly analyzed the problems of existing noise detection in dealing with actualsurveillance video algorithm perceptual noise. Thirdly, the paper described the birth anddevelopment of rough sets and rough entropy theory. Based on rough sets and roughentropy on video motion region, the paper provided removal algorithm that focus on thedistribution of the global properties and the locality of movement area of noiseperceptual, to solve the adverse effects of video motion area for noise detection. In theend, facing the problem of too bright or dark local areas interfer with noise detection,the paper provided an adaptive mechanism punishment,making full use of the statisticalproperties of the video image, deftly avoided the masking effect caused by the region.Furthermore, the paper proposed a blind detection noise perception index with theinformation combined with the concept of entropy. Through experiments, the paper verified the effectiveness of various methods in thepaper and meet the demands of current security monitoring system of blind detection onperceptual noise. These studies will make a useful exploration in the perception of noisedetection methods from camera surveillance video.