Detection of Respiratory Rhythm from Photoplethysmography Signal Using Morphological Operators

A new approach using morphological operators was proposed to detect the respiratory rhythm from the photoplethysmography (PPG) signal. In the study, photoplethysmograms were obtained from 5 healthy adult volunteers when they respired 6, 10, 15 times per minute. The reference respiratory signal was obtained by transthoracic impedance method simultaneously. Each PPG signal was processed using morphological operators to reduce the low frequency baseline drift and extract the peak envelope of the corrected signal first. Then the algorithm detected trend change of the pre-processed signal and obtained the rhythm of respiration. The result with a false rate of 4.52% showed that our technique had a good performance on detection of respiratory rhythm from PPG signal. The low computational complexity of the algorithm may make it easy to be implemented on MCU for real-time processing. More experimental data is necessary to improve the reliability and robustness of the algorithm.