Design And Implementation Of Video Surveillance Image Quality Diagnostic System

In recent years, with the innovation of video surveillance technology, high-definition digital video surveillance systems based on IP digital network has gradually replaced the original standard-definition analog video surveillance systems. Video surveillance of the major manufacturers are continue to offer the high-definition IPC with two million pixels and 5 megapixel HD NVR and various management systems and video surveillance platform. In this environment, innovation technology and video surveillance are no longer limited by the traditional front-end transmission distance. The distributions of the camera become more decentralized, the applicative range of video monitoring will be used wider and wider. Video surveillance system has now gone from original and some small LAN to metropolitan area networking, some sectors or units have achieved national network video surveillance system at all levels, and have formed a large WAN throughout the whole industry. However, the growth of video surveillance and the anomalies of video surveillance image quality make people found that the traditional work patterns are not keeping up with the growth of video surveillance.Recognition and diagnosis of the traditional video surveillance image can determine malfunction personally by depending on manual evaluation form, through the inspection of each image regularly or irregularly. This traditional model will cause huge image recognition and diagnosis, investigation and omission rate, process complexity, maintenance cycle increases when applied to large-scale video surveillance network, the workload but due to personal subjective factors in the investigation process, the image quality of recognition the diagnostic accuracy of a series of problems, will lead the overall image quality of video surveillance systems severely degraded, dramatically reducing system effects.Based on the above situation, this thesis will design and implement an automated diagnostic system of video surveillance image quality. The system can replace traditional manual evaluation mode, automatic identification of common abnormalities video quality through a certain algorithm, such as video image dithering, image color cast, blurring, and the system can record and feedback problems camera for users late maintenance and overhaul to provide fast, intuitive statistics. Through in-depth research study of subjective and objective video quality evaluation method, combining the features of video surveillance images and video surveillance system maintenance characteristics of camera shake and image blur image color shift associated algorithms and the algorithms and diagnostic The results of the statistical analysis done. Meanwhile, this paper based on the needs of video surveillance image quality diagnostic system analysis, the system is divided into diagnostic set, pre-processing video frames, video quality and diagnostic results are stored in five sub-function modules, and more than four sub-function modules for the design and realize, after a lot of experiments to test the system and achieved good results, in line with expectations design goals.