Research On Detection Method Of R-peak Of ECG

At present,the number of patients with cardiovascular disease continues to grow,and cardiovascular disease has been a serious threat to human health.In medical clinic,doctors diagnose angiocardiopathy based on ECG.So how to extract useful information of ECG has become the focus of our research.ECG is a weak biological signal,which is easily drowned by noise.So before we do ECG research,we need to filter the signal.After the signal is preprocessed,the accurate location of R wave is another key issue in the study of ECG.If we can't solve these two problems,we can't make accurate ECG automatic analysis,and doctors need to spend a lot of energy in ECG analysis.Manual electrocardiogram analysis c onsumes a lot of resources and can not give real-time diagnosis results,which is not conducive to the treatment of patients.Therefore,this thesis focuses on the two problems of filtering and preprocessing of ECG signal and how to locate the location of R peak.The last three chapters of this thesis are arranged according to the three modules of our algorithm: signal preprocessing,R wave feature extraction and R peak detection.This thesis mainly studies the following aspects:(1)A preprocessing algorithm based on EMD and permutation entropy is proposed.The proposed algorithm can simultaneously filter the high frequency interference noise and low frequency baseline drift noise.The algorithm can retain more details of the ECG signal at the same time.Finally,we use permutation entropy to select IMF for reconstruction signal,which can avoid the influence of IMF's subjective consciousness.(2)The method of curve length transformation is used to deal with ECG signal,which can effectively enhance the R peak characteristic of ECG signal,and greatly reduce the situation of signal misdetection and leakage detection.(3)Physiological ECG knowledge is used to improve the amplitude adjustment algorithm so that the algorithm is more suitable for the actual situation of ECG.Moreover,additional detection conditions are added to the algorithm,so that the algorithm can detect the location of the R peak and reduce the missed detection rate effectively.(4)In this thesis,we use the proposed algorithm to test the data of the entire MIT-BIH database.The detection results are as follows: the sensitivity is over 99.85%,and the positive detection rate is over 99.03%.The algorithm studied in this thesis can retain most of the characteristics of ECG,and it can also effectively filter noise.And the location of the R peak can be detected accurately.These studies can provide an important basis for the diagnosis and analysis of clinical electrocardiogram.