A Novel Real Time Baseline Wandering Cancelation in ECG Signal Based on Morphological Filters

Electrocardiogram (ECG) signal is an important signal which shows electrical activity of the heart. It should be noise free for analyzing and extracting true information. The aim of this study is to remove the baseline wandering noise. This is a low frequency noise that makes distortion in ECG signal and sometimes makes it unusable, because extracting valuable information is impossible when this noise exists. We first introduced common methods and their limitations that are used for cancelation of this noise. Then a novel method based on morphological filters of Image processing with a specific structural element is proposed. In this method an estimation of baseline wandering noise is obtained by dilation and erosion, and then it is subtracted from noisy signal. Results confirm that the output signal is noise free without any distortion in vital part of ECG. Also a simple threshold is used to detect the locations of R peaks of de-noised signal and the result is compared with Pan-Tompkins method, which the thresholding after de-noising method was more accurate than Pan-Tompkins. The implementation of this algorithm is easy and fast in response, so it can be used both in real time and offline data, and because of its fast response, the portable devices can use this algorithm to make their output signal noise free.

کلیدواژه ها:

Electrocardiogram (ECG) ، Baseline Wandering noise ، Morphological . Filters