Design And Implementation Of Face Detection In Target-driven Surveillance Video Positioning System

Traditional monitoring activities identify the monitored objects manually.These video surveillance systems have a single function and can only store and play back video.In the artificial real-time monitoring method,the monitoring personnel are easily fatigued.When the number of monitoring videos is large,it is easy to omit important information and it is difficult to accurately find a specific target.Based on the above reasons,surveillance activities urgently require intelligent video surveillance systems to help the monitoring personnel to better complete the work of finding all personnel in the surveillance.For these phenomena and problems,this paper presents the design and implementation of the face detection method in the target-driven surveillance video positioning system.The face detection method in this paper has the advantages of high accuracy and low computational cost.We design and implement a target-driven surveillance video positioning system in this thesis.The system is divided into four functional modules:video processing module,face detection module,face recognition module and camera rotation module.In the video processing module,we cut the video stream collected by the camera.The obtained image is processed by the face detection module and the face recognition module,and the position information of the specific target in the image is found.Then we issue a command to the camera in the camera rotation module for a specific target-driven monitoring activity.In addition,this thesis focuses on the application of face detection method in this system,designs a convolutional neural network structure and a Latent SVM classifier,and combines them to realize the face detection method in this system.Finally,we carry out functional testing and performance testing of the target-driven surveillance video positioning system in this thesis.The experimental result shows that the system has achieved its basic function of detecting a specific target and of continuously monitoring it as a driver.The video processing module,face detection module,face recognition module and camera rotating module in the system all can operate normally and realize the function of this module.In addition,the face detection method in this thesis has a good performance on the important evaluation criteria of face detection such as detection rate,false detection rate and missed detection rate under the premise of ensuring a certain real-time performance.