Research On Target Recognition And Security Application Based On Deep Learning

Vision is an important channel for human beings to communicate with the outside world.Computer vision captures,recognizes,and understands objective information by simulating the form of human vision,and continually improves people's lives.The development of computer vision technology has promoted the large-scale application of video surveillance systems.However,the traditional video surveillance system requires great human and material resources to maintain,and lacks real-time performance.It can no longer meet the development needs of video surveillance systems,and security surveillance systems continue to develop in an intelligent direction.The intelligent video security monitoring system is an automatic processing process of video information using computer vision technology.The intelligent security system does not need to be manually guarded,and uses computer vision and image processing technology to automatically process and analyze image information.With the development of deep learning and computer vision,people put forward higher requirements for the function and quality of security monitoring system.Intelligent video security system has shown great application prospects with its intelligence and real-time performance.This paper mainly researched how to use the embedded development board to integrate the external equipment to build an intelligent video security monitoring system,which opens up a new way for the development of intelligent video security monitoring system with small size,low price and stable performance.This article is based on the Raspberry pi and OpenCV,design a intelligent security system based on deep learning.This paper first analyzes the shortcomings of traditional video surveillance systems and the research status of intelligent video surveillance systems,summarizes the current needs of intelligent video surveillance,and expounds the necessity of research on intelligent video surveillance systems.Then the idea and method of moving target detection technology are studied.Combined with the characteristics of embedded platform,a design scheme of intelligent security system based on deep learning is proposed.The overall framework and various functional modules of the system are analyzed and designed.Finally,the intelligent video security system was tested and the design scheme was summarized.The corresponding deficiencies and improvements were proposed.