The Research Of Acquisition And Analysis Of ECG Based On The Virtual Instrument

With the ascending tendency of the global population aging, how to properly resolve the old-age security problem of the old people in the empty nest families will become an important subject that human have to face. Faced with this social problem, it's urgent to develop a home sevice robot, and multi-parameter monitoring system is an important research part of home service robot. Electrocardiogram (ECG) contains the most basic and important physiological information of the human body and it is an important basis for diagnosis of cardiovascular disease, furthermore, ECG technology has characteristics of simple, noninvasive, easy to read, good repeatability and so on, so ECG becomes the main monitoring content of the multi-parameter monitoring system.As a sub-project of the National High-Tech Research and Development Program "Home Service Robots in Intelligent Space", in this paper, we design a suit of acquisition and analysis system of the twelve lead ECG signals, which is based on virtual instrument, and at last the collected signals and the preliminary analysis results are sent to the robot. The main contributions of this subject are as follow:(1) Because the USB bus has the advantages of Plug-and-Paly, simple interface and the high transport rate, this paper chooses USB data acquisition card to collect the twelve lead ECG data. The USB data acquisition card, which is come from Beijing's Altai, is directly connected to the USB interface of the computer, and applications are writed by calling the dynamic link library in the Labview platform, in this way, we can transmiss data between the PC and the lower computer.(2) Because the ECG is extremely weak and easily affected by background noise, the pre-processing algorithm we used is based on wavelet transform. We improved the existing threshold processing method, determined the optimal wavelet and the best decomposition level, and then processed wavelet coefficients of different level with different thresholding methods, removing the main noise of the ECG, such as the baseline drift, the frequency interface and the muscle interface, purposeful.(3) After pretreatment in the signal, this paper proposes a single lead based feature extraction method, and the Mallat algorithm is improved by removing the second sampling. We use trous algorithm with wavelet transform. In the process, the modulus extremes of the wavelet coefficients are looked for, in order to locate the R-wave. In addition, a range of compensation strategies are used to improve the accuracy of detection, and finally the algorithm is evaluated with the data of MIT-BIH Arrhythmia Database.(4) As a single lead based features extraction method has some drawbacks, this paper proposes another method, and we use wavelet entropy to extraction features with multi-lead singles. We reduce the dimension of the signals by principal component analysis, and then wavelet transform is used to the processed signals. In order to determine the locations of R-wave, the signals are classified by the peak number and variance of the wavelet entropy. Finally the algorithm is assessed with the data of St. Petersburg arrhythmia database.(5) A twelve-lead ECG acquisition and analysis system is designed. We prepare the graohical user interface, real time acquisition and display module, data playback module, data analysis and processing module and data memory module. In the data memory module, two data stored method are used, one stored method is based on the database, and the other is based on the XML document.