Human Life Feature Detection Based on EMD Method in THz Band

Non-contact respiration and heartbeat signal detection could be applied to search for victims in the disaster or the remote monitoring of the respiration and heartbeat of a patient. Considered the human respiration and heartbeat signal is weak signal compared to the complex environment, we proposed a human life feature detection method based on EMD in THz band. A human target echo model for THz radar was established firstly, and a time-frequency analysis of the echo after empirical mode decomposition will yield a signal. The centroid curve of this signal underwent a second time-frequency analysis to extract and separate the heart and breath rates of the individual. The simulation carried out in Gaussian clutter background show that the detection method based on EMD has a better performance than the direct detection method. It has the strong ability to resist noise and is suitable for weak signal processing.