Implementation Of The Face Recognition System Based On Intelligent Video Analysis

In recent years,the computing speed and computing power of artificial intelligence chip have been greatly improved,which provides a good hardware platform for the transplantation of deep learning algorithm in video surveillance system.Based on the theory of high-speed circuit design,this paper designs a face recognition system based on Hi3516DV300 by using intelligent video analysis technology and combining with practical application scenarios.On the premise of meeting the basic monitoring function,it realizes the front-end deployment of face detection function based on deep learning algorithm.At the same time,the face contrast recognition algorithm is deployed in the video server web master.The main work of this paper is as follows:Firstly,the hardware implementation architecture of the system is analyzed in detail.The main parameter specifications of the system are clarified and the hardware composition of the system is designed;Secondly,the generalized signal integrity and other high-speed circuit design theories and printed circuit board engineering design have been deeply studied.Based on the transmission line theory,the theories and applications of signal reflection,crosstalk,power integrity and electromagnetic compatibility are studied;Thirdly,the principles of the system's image acquisition subsystem,storage subsystem,and power supply system are analyzed in detail,and the hardware design of the system is completed.At the same time,the power ripple and electrostatic immunity tests are performed on the entire hardware system to verify the system design;Finally,I learned the face recognition algorithm and model transplantation based on the Caffe1.0 deep learning framework.After a lot of engineering practice,the reliable operation of the hardware system has been realized,and the design requirements of the face recognition system have been completed.