Life signal feature extraction method based on line spectrum tracking

Abstract

The invention discloses a life signal feature extraction method based on line spectrum tracking. The life signal feature extraction method includes the steps that modeling is performed on a target chest motion; N frequency-modulated signals are continuously transmitted by a radar, frequency mixing is performed on echo signals to obtain an IF signal and obtain a target phase, and a phase vector phiof length N is formed; STFT is performed on phi to obtain a time-frequency spectrum matrix P<s>; a first HMM model gamma <1> is constructed; the P <s> is divided into Q groups, each group of data isused as an observed quantity sequence, and a Viterbi algorithm is adopted to calculate an optimal state sequence corresponding to the gamma <1> observed quantity sequences to obtain frequency units inwhich the breathing frequency is located at each moment; elements of multiple harmonic components, corresponding to the breathing frequency, in the P<s> are set to be zero, a second time-frequency spectrum matrix P'<s> is obtained, and a second HMM model gamma <2> is constructed; and the P'<s> is divided into Q groups, each group of data is used as an observed quantity sequence, and the Viterbi algorithm is adopted to calculate an optimal state sequence corresponding to the gamma <2> observed quantity sequences to obtain frequency units in which the heartbeat frequency is located at each moment. According to the life signal feature extraction method based on line spectrum tracking, the breathing frequency and heartbeat frequency of a target can be accurately extracted without contact.

Description

technical field [0001] The invention belongs to the technical field of signal processing, and in particular relates to a feature extraction method of vital signals. Background technique [0002] Respiration and heartbeat are important indicators of vital sign information. On the one hand, cardiopulmonary sign information can be used to judge whether there is a living body, and the basic state of the living body; on the other hand, abnormal cardiopulmonary activity parameters are often accompanied by medical emergencies. Therefore, real-time monitoring of cardiopulmonary signs has very important practical value in many occasions. Among the common vital sign signal detection methods, the methods used for human respiration detection mainly include: pressure sensor method, temperature sensor method, electrical impedance respiration measurement method, respiratory induction plethysmography method; detection methods related to heartbeat include: electrocardiogram, Heart sound, p...

AI Technical Summary

Problems solved by technology

[0003] Most of the above methods are based on contact, and need to be in contact with human skin to measure vital sign parameters, which limits their application in some special occasions

For example, in the monitoring of the elderly, wearing equipment for a long time will cause discomfort to the monitoring object; for patients with extensive burns, electrodes or sensors may cause secondary damage; in the search and rescue process after some disasters, through contact It is obviously impractical to detect vital signs by means of a method, so the development of non-contact vital signal feature extraction technology has very important value

Images