Research On Detection And Prediction Of Respiratory Signals

The information about the rhythm,intensity and frequency of the human respiratory signal can reflect the pathological changes of human respiratory organs to a great extent.Therefore,it is of great medical significance to monitor human respiratory status.However,at present,the medical respiratory monitoring equipment is often bulky or expensive,and more direct contact with the sensor and the human body contact detection method,will bring inconvenience and restraint to patients At the same time,some of the activities of the human body may cause the sensor to fall off and move,making the test results inaccurate,and even failure detection.Aiming at the above problems,this paper research on how to access to human respiratory signal accurately and stably under the condition of avoiding direct contact with the human body,regarding human respiratory movement as the research object,using the camera as a sensor,combined with image processing and signal processing technology.According to the characteristics of the correlation between the human respiratory motion and the undulating motion of the human body abdominal or chest,this paper presents a method of non contact type respiration detection based on motion amplification in order to improve the accuracy of respiratory measurement.A receiver is used to receive the light irradiated on the human’s abdominal or chest reflected from a laser pen,the light spot image on the receiving device is collected via webcam,and the image motion tracking algorithm is adopted to track the movement of the spot automatically,so as to obtain the respiratory signal of the human bodyThe image processing technology is used to detect and track the light spot trajectory on the receiver to characterize the human body breathing movement,so as to indirectly obtain the body’s breathing signal.According to the characteristics of high brightness light spot,this paper proposed an adaptive threshold detection algorithm,starting from the highest gray level of the gray level histogram of the light spot image to traversal,finding the gray level of the corresponding points of the first trough as threshold.Using the algorithm to conduct comparison experiment on threshold segmentation on spot image with Otsu method and iterative method.The experimental results show that the algorithm can accurately detect the light spot.After detecting the light spot successfully,according to the characteristics of light spot movement,atracking algorithm based on Kalman filter and adaptive template matching is proposed.The tracking algorithm is used to conduct comparison experiment on the track the light spot in the image sequence with template matching tracking algorithm.The experimental results show that the proposed tracking algorithm has the advantages of high real-time and high accuracy.In order to improve the accuracy and robustness of respiratory detection,a method of spatial constraint was proposed to solve the interference caused by non respiratory movement such as cough,body shake and speech.For the lag problem of respiratory signal detection caused by the time occupation of the image acquisition,image processing and signal processing process,using signal prediction algorithm based on adaptive filter for respiratory motion as compensation,respiratory signal detection method is proposed in this paper which can adapt to the situation of high real-time breathing test requirements.Based on the theoretical study of respiratory signal detection and respiratory signal prediction,the corresponding respiratory signal detection system is realized.The system was used to detect and predict respiratory signals in different volunteers.The experimental results show that the respiratory signal detection method proposed in this paper can amplify the characterization of human respiratory small chest or abdominal movement,so as to improve the accuracy of detection of respiration.At the same time,this adaptive respiratory signal prediction algorithm used in this paper has the characteristics of high precision and strong robustness.