Surveillance Camera Interference And Image Anomaly Detection System

Video surveillance system is an integrated system of prevention ability, in recent years, with the increasing awareness of people’s security, video surveillance systems have penetrated into all areas of national life, banking supervision, residential security monitoring, traffic violation monitoring, etc. With so many camera applications, to detect whether each camera monitor screen is normal, will be very difficult.Therefore, we designed and implemented a Surveillance camera interference detection system that can automatically detect the captured images to determine whether the picture is abnormal and human disturbance.Surveillance cameras are in the front of video monitoring system. They are easy to be disturbed. Especially some suspicious persons in order to avoid monitoring, usually make various interference action on the camera, including camera rotation or lens shade and so on. Related studies found that a significant change in image characteristics may occur after the interference, based on these studies we can using SIFT image feature functions to obtain characteristics of the image by transformation the image feature to determine whether the camera interference occurs, and combining PCA-SIFT to the image feature vector dimension reduction, reducing the amount of calculation and improve efficiency.In addition to human disturbance, the surveillance camera itself is also affected by the surrounding environment, network transmission, hardware failure. It produce an image anomalies which include static frame, black, mosaic, color cast, and so on. Resulting in loss of effective information. Through the research on the quality of video detection, we designed and implemented a surveillance camera picture anomaly detection scheme which can detect static frames, black, mosaic, color cast, and can immediately issue an alarm.Video surveillance system may also appear fuzzy picture, abnormal noise and other issues. Through the analysis and comparison of various video quality assessment method, we designed a no reference objective video quality evaluation model.