Research On Smart Home System Based On Human Behavior Recognition Algorithm

With the rapid development of science and technology,intelligent home is gradually integrated into people's daily life.Human behavior recognition based on machine vision,smart home products have also quickly entered people's vision.Most of the traditional smart home products are mainly manual operation and automatic control,which lack flexibility.In view of these current situations,this paper mainly studies the behavior recognition algorithm based on human gesture recognition and applies it to the smart home system.The design and production of hardware circuit and software system show that the studied algorithm can effectively capture human gesture movements,and the designed software and hardware system can realize timely and effective response to different gesture movements,So as to realize the gesture control function of different household devices.In view of the current situation that the gesture recognition algorithm is difficult to transplant to the system,combined with the research status of gesture recognition,this paper proposes an improved gesture recognition algorithm,which aims to improve the recognition accuracy and reduce its complexity,so as to ensure that it can be implemented in the built software and hardware system.Considering the stability,maintainability,portability and cost performance of the system,this paper has done a lot of research on gesture recognition algorithm and software and hardware design.The main work is as follows:1.Aiming at the problems of high complexity of gesture recognition and sparse characterization of image,this paper studies a gesture capture algorithm based on moving target and an image sparse characterization method based on regularization to improve the accuracy and efficiency of gesture recognition,and preliminarily demonstrates the effectiveness of the algorithm through simulation.2.After investigating the existing human behavior recognition smart home system on the market,combined with the existing experimental conditions,a set of efficient,reliable and cost-effective smart home software and hardware system is selected and designed,and the interference between modules and system compatibility are fully considered,so as to carry out the circuit design.3.Based on the theoretical analysis,a host computer system with multi-functional data management is designed on the computer side,which can realize the functions of video acquisition,gesture capture and recognition.The wireless network established by Zig Bee is used to realize the data interaction with the hardware system,so as to realize the gesture control function in the smart home environment.After the fusion design of the algorithm and the software and hardware system,the function of the system is also tested and verified.The results show that the hardware function modules designed by the system can work normally,the gesture recognition accuracy of the designed algorithm is high,achieves the expected effect,and can effectively realize the control function of smart home.