| Pulse signal contains a wealth of physiological and pathological information of the human body, which is an important parameter to characterize the physiological condition of the human body. In order to promote the objectivity of pulse diagnosis, a binocular vision pulse image acquisition system was developed. The system obtained the pulse signal of radial artery based on the principle of visual measurement.However, due to the combined effects of various factors, there are errors in the human body pulse signal obtained by visual measurement, which have a great influence on the accuracy of subsequent pulse analysis and pulse information extraction. Therefore,it is necessary to analysis the accuracy of the system in order to carry on the calibration of the system. To solve this problem, a precision analysis method of binocular vision pulse image acquisition system based on laser displacement sensor is proposed, which mainly completes the following work.(1 )The experimental system of precision analysis is designed and implemented.The pulse image sensor based on the principle of space displacement measurement of"best cutting pulse pressure” and thin film stress deformation is improved, the control system of synchronous trigger laser displacement sensor and dual camera is designed.This paper introduces the design process of the control system, the trigger principle and the running state, and by improving the pulse image sensor, the field of view of the camera is enlarged, the laser displacement sensor shares the same field of view with double camera.(2) The vertical displacement of the feature points on the thin film probe was extracted. The pulse is obtained by camera calibration,image noise reduction,edge feature extraction, ellipse fitting, center of mass acquisition and calculation of pulse three- dimensional information.(3) The vertical displacement of the feature points on the thin film probe was extracted from the laser displacement sensor. The noise of the acquired pulse signal is analyzed to remove the noise such as optical interference, baseline drift and power frequency interference during data acquisition. According to the characteristics of the noise, the Butterworth digital low-pass filter is used to denoise the signal to obtain accurate pulse signal.(4)The accuracy of binocular vision pulse image acquisition system is analyzed.By analyzing the data of 56 sampling points on the thin film probe, the measurement accuracy of the system is obtained. Firstly, the pulse displacement sensor is analyzed by the correlation between the pulse signal peak-to-peak value and the vertical displacement of the pulse obtained by the dual camera, the sampling point with large influence on the random interference is removed. Secondly,the weighted average algorithm is used to calculate the pulse signal peak-to-peak value, so the value of the pulse signal peak-to-peak of the corresponding sampling point can be obtained. Then the longitudinal displacement error of each sampling point is calculated to obtain the accuracy of the system. Again, the systematic error characteristics are studied. The strong correlation point is divided into test set and verification set. The error of the two sample sets is fitted by the least squares polynomial curve, and the results of the two sample sets are compared and analyzed. It is proved that the error characteristics obtained by the fitting of the test group meet the error characteristics of the whole system. At last, the system error is summed up, and the system error increases gradually with the weakening of the pulse.By analyzing the accuracy of the system, it provides a basis for the follow-up analysis of physiological and pathological information of human body through pulse analysis. |
