ECG Signal Intelligent Processing Key Technology Research

ECG(Electrocardiogram,ECG)signal is a kind of human physiological signals,it can provide some heart and disease information,and therefore is widely used to detect whether the heart health.However,during the signal acquisition process,ECG signals are inevitably subject to various interferences,including baseline drift and random noise.Baseline drift of ECG signal is mainly caused by the noise generated by the tester’s breathing or movement,which is a common problem and cannot be avoided.Its low frequency overlaps with the ECG signal’s ST segment,elevating the ST segment and compromising judgment.At the same time,ECG signals are difficult to separate from noise due to their weak ECG signals,which are easily affected by ambient noise,and the spectrum of noise overlaps the ECG signal.Therefore,the processing of ecg signals has become a key and difficult problem.This paper will design and apply ECG signal processing algorithms and make the following innovations:1.The mechanism of ECG signals and the components of ECG signals were studied in detail,and the sources of noise in the ECG signal acquisition process were analyzed,and the sparsity characteristics of ECG signals and differential signals were further studied through simulation experiments.2.In the process of noise reduction of traditional ECG signal filtering algorithm,due to a large amount of compression of the peak value of the original signal,the ECG signal peak value after processing is under-estimated and the original ECG details cannot be recovered,etc.,a noise reduction and baseline correction algorithm based on the sparse characteristics of ECG signal is proposed.For ECG signal with noise and baseline drift,noise reduction and baseline correction at the same time,a lot of experimental simulation is carried out by MIT-BIH database,and compared with traditional filter algorithm has carried on the omni-directional,results show that the peak owe estimation problem in traditional filtering algorithm,effectively improve and maintain the ECG waveform characteristics of the original signal.3.The algorithm proposed in this paper used a large number of sparse banded matrices and high order non-convex optimization functions,which greatly saves the running time.The choice of penalty term and difference order is studied and analyzed in detail.First of all,based on the symmetric penalty term for signal processing,further,through the analysis of the characteristics of ECG signal using a composite punishment to ECG signal processing,namely the original ECG signal is asymmetric punishment,punishment and its differential signal is symmetry achieved the effect of further accurate punishment,make to the original ECG signal estimation is more accurate.4.In 12 lead wearable ECG monitoring garment on application of LTI sparse filter and noise reduction and baseline correction algorithm,test,obtained the satisfactory filtering and noise reduction effect.The application value of the proposed algorithm is verified by comparing the normal ECG signal with the ECG signal collected and displayed after filtering by this algorithm.The overall design of the algorithm in this paper achieves the expected goal and has certain practical and application value.