Development Of Artificial Intelligent Surveillance System Based On Deep Learning

As is well known,surveillance systems have extremely important application value in the fields of society,transportation,military and public security services.With the rapid development of artificial intelligence technology,the research and development of new intelligent visual surveillance system have become a research hotspot in the application of artificial intelligence.Compared with the traditional surveillance system,the artificial intelligence surveillance system can not only make the surveillance personnel free from the complex and massive video information,but also make the surveillance function more automatic and intelligent.In recent years,deep learning has made breakthrough in the field of computer vision and pattern recognition,and has become one of the important technologies in the application of artificial intelligence.To this end,this paper aims to investigate and develop an artificial intelligence surveillance system based on deep learning theory.The main work of the paper is as follows:1.Based on the analysis of the demand of artificial intelligence surveillance system,the technical scheme of the system is put forward.The scheme consists of object detection and tracking as well as action recognition based on deep learning.2.The object detection algorithm based on deep learning is studied.This paper mainly studies the object detection algorithm R-CNN based on candidate regions and the object detection algorithm YOLO based on regression.An object detection and tracking model based on YOLO network and kalman filter is proposed in order to ensure that the intelligent surveillance system can process the transmitted video stream in real time.The model uses YOLO network to quickly detect the target object in the current frame,and predicts the position information of the object by kalman filter tracker.The predicted position information matches the position information of the object detected in the next frame,so as to achieve the purpose of rapid tracking.Experimental results show that the proposed model can be applied to real-time surveillance of intelligent surveillance system.3.A action recognition algorithm based on deep learning is studied.Particularly,the two-stream network structure and C3 D network structure in deep learning are studied.In order to ensure the recognition accuracy,the system can quickly identify the motion.A motion recognition model based on C3 D network is proposed.The experimental results show that the proposed model has high recognition accuracy on the self-made database and can be used for the actual surveillance of the system.4.The whole realization process of an artificial intelligence surveillance system is introduced including the collection and construction of databases,the fusion of algorithm models,the improvement and treatment of practical application problems,and the application results of the final system.