Research On Recognition Method Of Gesture And Action Needs Of Elderly Disabled People

Since 21 st century,with the severe aging problem,the number of elderly people characterized by slurred speech and limited mobility is increasing.How to accurately identify the nursing needs of such people has become a key issue to be solved in the field of old-age care and medical care.In recent years,gesture recognition technology has become an important breakout to solve the above problems due to its significant breakthrough in the theory of algorithms.Therefore,based on the research of gesture recognition algorithm and the gesture behavior habits of elderly disabled people,this paper establishes a custom sample database by volunteers and designs a gesture recognition human-computer interaction system based on nursing bed.The work done in this article can be divided into the following four parts:(1)The preprocessing process of gesture image based on vision is studied.Based on the analysis of the advantages and disadvantages of different color spaces under the skin color model,the original image is enhanced.The algorithm of skin color region segmentation for Tangos model under complex background is given in detail.Finally,morphological filtering and face elimination algorithm are used to obtain the hand region.(2)Two different methods of gesture feature extraction,static and dynamic,are analyzed.Aiming at the feature extraction of static gesture,on the basis of studying the feature extraction method of static hand profile,the characteristics of three kinds of static gestures are filtered and fused in combination with the actual needs of this paper.Aiming at dynamic gesture feature extraction,two gesture tracking algorithms,camshift motion tracking and pyramid optical flow tracking,are compared,and the dynamic gesture trajectory tracking of video images is realized according to the image optical flow.(3)The static and dynamic gesture recognition methods are studied.This paper introduces the classification recognition of custom static gesture sample library based on support vector machine.This paper presents the fusion of dynamic optical flow vector trajectory characteristics with static hand-shaped features,and uses dynamic time regularization method to identify gesture feature sequences,which effectively solves the problem of the same trajectory gesture recognition with different hand shapes.(4)A single chip microcomputer nursing bed control system based on STC15F2K60S2 is designed.Combining with the WiFi communication module,the static gesture recognition algorithm studied in this paper is applied to the attitude control of the bed body,and thedemand recognition based on the user’s static and dynamic gesture behavior habits is completed,and the system test is carried out.