| In recent years,overuse of Visual Display Terminal(VDT)has led to an increasing number of patients with ophthalmic diseases.Research on visual fatigue monitoring methods is of great significance and value to reduce the damage caused by overuse of VDT.In this paper,the relationship between heart impact signal and VDT visual fatigue is studied in order to extract visual fatigue index and design visual fatigue self-determination system.The main work is as follows:(1)The method of monitoring VDT visual fatigue using heart beat signal is proposed in this paper.Firstly,according to the principle that the correlation coefficient between acceleration signal and ECG is the highest,the seating posture of the seat and the tester is determined.Then,VDT visual fatigue experiment is carried out,and acceleration signal obtained is de-noised twice to eliminate the non-linear noise interference such as baseline drift,human random motion and ground vibration,so as to obtain the acceleration signal in the range of ECG frequency.Finally,inverse Fourier transform is used to extract heart beat signal,and the correlation between the heart beat signal and synchronously collected ECG is studied.The linear correlation coefficient is 0.9918 and the non-linear correlation coefficient is 0.796.The conclusion that heart beat signal replaces ECG signal for VDT visual fatigue monitoring is verified,which lays the foundation for unconstrained VDT visual fatigue monitoring.Foundation.(2)The activity of sympathetic nerve can reflect the degree of visual fatigue,and the low-frequency value of heart rate variability(LF)can also reflect the activity of sympathetic nerve.LF is used as a judgment index of visual fatigue in this paper.After calculating RR interval of ECG signal and JJ interval of heart beat signal by using wavelet algorithm,the LF values of both signals are obtained.Then LF curves are drawn,and the conclusion that the trend of LF curves of both signals is consistent is drawn.The subjective evaluation based on questionnaire is made by using SD subjective evaluation method,and the accuracy of judging visual fatigue by using LF is verified by the consistency of confidence intervals.(3)VDT visual fatigue is classified according to LF threshold.Four kinds of kernel functions of support vector machine are used to study and judge 100 sets of visual fatigue data respectively.It is concluded that Polynomial kernel function has the highest recognition accuracy of VDT visual fatigue,which is 83.33%.Finally,the VDT visual fatigue judgment and recognition system based on support vector machine is designed and verified.The overall recognition accuracy can reach 93.75%.This proves the feasibility of the support vector machine to discriminate visual fatigue,and also proves the feasibility of the VDT visual fatigue monitoring method based on unconstrained heart impact signal extraction. |
PDF Full Text Request