Development Of Human Sitting Posture Intelligent Recognition System

In recent years,human health problems caused by sitting posture have become increasingly prominent,and prolonged sitting has been listed as one of the ten most fatal and disease-causing killers.The research shows that the sitting posture problem can be effectively avoided by feeding the sitting posture status to the user in real time.Therefore,this paper conducts a series of researches on the human sitting posture intelligent recognition system.In order to divide the sitting posture reasonably and formulate the sitting posture sample collection strategy,the sitting posture information acquisition module is developed based on the collection requirements,and a reliable sitting posture recognition algorithm is proposed according to the sitting posture information characteristics.software.The research content of this subject can realize the online rapid collection and accurate identification of sitting posture information,timely remind bad sitting posture and sedentary behavior,and provide sitting posture health assessment,which is helpful for people to avoid the harm caused by sitting posture to physical health.The specific research work is as follows:(1)The overall scheme design of the sitting posture intelligent recognition system is carried out.The system design ideas were sorted out,the basic research on system design was carried out,and the overall structure of the sitting posture intelligent recognition system was designed.(2)The sitting posture information acquisition module is designed,and a real-time optimization mechanism is proposed,which can achieve a wireless acquisition efficiency of 0.404s/time for a 60×60 piezoresistive array.This module uses a highperformance single-chip microcomputer as the main controller to realize cross decoupling and voltage division sampling,optimizes various real-time influencing factors,designs multi-column synchronous gating circuit,realizes line-by-line cyclic triggering,multi-column parallel acquisition mechanism,improves the Acquisition efficiency;theoretically analyze the RC filter circuit,propose a resistor-capacitor selection scheme to achieve a balance between speed and accuracy;propose a framing format based on a limited force area,reduce the amount of wireless data,and achieve rapid collection of sitting posture information.(3)A CNN-SE-ELM sitting posture recognition algorithm model is proposed,which realizes the reliable recognition of sitting posture.Row filling and adaptive median filtering are used for data preprocessing,and accuracy,recall and precision are used as evaluation indicators for classification algorithms,and different algorithm models are compared and analyzed on each sample set,and the algorithm model proposed in this paper is verified.The results show that the CNN-SE-ELM algorithm can achieve a sitting posture recognition rate close to 95% and 90% on the single and group(18 people)sample sets,respectively.(4)Developed the sitting posture intelligent recognition software.The sitting posture was monitored online at a frequency of 2s/time,and the sitting posture state during the monitoring period was comprehensively evaluated according to the sitting posture quantitative index and the sitting posture health assessment model.The research results of this topic can be applied to the development of intelligent equipment for human sitting posture health monitoring,which can collect and identify users’ sitting posture status online,and provide reminder and health assessment functions,which can help people correct bad sitting posture in time and improve their sitting posture in a targeted manner.question.