Research On Security Behavior Management Detection Technology Based On Video Surveillance

Nowadays,with the continuous development of science and the advancement of science and technology,the development of video surveillance systems has gradually grown,and has involved various occasions of human life.But now most video surveillance systems are still in the traditional mode,that is,only the record is not judged.At present,security personnel are mainly used to continuously monitor sudden and suspicious behaviors that may occur randomly.In this process,a large amount of manpower and material resources are required for safety management,but the human body will feel tired and will not continue to pay attention.This makes it easy to cause a missed inspection and loses the meaning of timely monitoring.The video behavior monitoring and management technology is an automatic identification process for abnormal and unsafe behavior.This technology research has great development prospects.It can not only reduce manual management,but also prevent accidents in advance.An important technical method in the research of video behavior monitoring and detection technology is to detect and identify the foreground target of video.The rational use of this technology will bring great convenience to human beings.It can automatically detect and identify video,find abnormal unsafe behaviors,and issue alarms for handling,which greatly facilitates humans.In order to achieve an unusual situation,the camera device can alarm itself,firstly extracting the captured image,and then identifying and discriminating the extracted image.At present,there are many scholars conducting research on video surveillance behavior recognition,but there are still many problems(such as object noise,cavity and background edge misjudgment when monitoring camera jitters).In order to solve some problems in the research of security behavior management detection technology for video surveillance,this paper proposes an improved video surveillance foreground object extraction algorithm and video surveillance anomaly behavior recognition algorithm.The main research work and innovations are as follows:1.Improved the video surveillance foreground object extraction algorithm.In order to improve the often used video surveillance foreground object extraction algorithm to extract the object from noise,cavity and monitor camera jitters,the background edge misjudgment,etc.,this paper proposes the extracted foreground object.The video is Gaussian filtered first,then the media is filtered,and finally the second resolution algorithm is used to reduce the foreground object holes and smooth the background edges.When the object is in a static background,the adaptive combination of the Gaussian model is used to obtain the desired object;when the object is in the dynamic background,the PBAS is usually used to obtain the desired object;when the camera is shaken,the second resolution algorithm is used.And the proposed improved method is used for comparative experiments.2.Improved the algorithm for detecting and detecting abnormal unsafe behaviors.Through video surveillance,personnel activities can be monitored in public places to prevent abnormal accidents.It is very difficult to continuously watch public places,so intelligent video surveillance is required to monitor human activities in real time and classify them into regular and abnormal activities,and can generate alerts.Accordingly,an algorithm based on feature descriptors encoding motion information and classification methods to effectively solve the problem is proposed.The new anomaly indicator comes from a hidden Markov model that learns to observe the histogram of the direction of the optical flow of the video frame.The indicator measures the similarity between the observed video frame and the existing normal frame,and can effectively detect and identify abnormal unsafe behavior in the surveillance video.Figure[20]table[0]reference[70]...