本文關鍵詞：容積脈搏成像特征提取方法研究及在生理信號檢測的應用　出處：《哈爾濱工業(yè)大學》2016年博士論文　論文類型：學位論文

  更多相關文章： 容積脈搏成像 特征提取 心率 脈搏波傳導時間 血氧飽和度

【摘要】：脈搏波蘊藏了豐富生命信息。通過對脈搏波的形態(tài)、強度、速度和節(jié)律的分析,可以獲得生理和病理特征。在環(huán)境光條件下使用攝像頭能提取容積脈搏波,它是借助容積脈搏成像(Photoplethysmographic imaging,PPGi)檢測血液容積變化的方法,通過計算PPGi圖像像素均值使二維信號轉化為一維信號可以獲得容積脈搏波。日益普及帶有攝像頭智能手機和移動式個人終端設備提供了一個低成本和泛在式的健康感知平臺,人們可以在非醫(yī)院、非約束的環(huán)境下使用移動終端設備長時間實時動態(tài)地評估個人的健康狀況,達到對疾病的早期預警、早期預防、早期發(fā)現(xiàn)、早期診斷和早期治療的目的�？梢灶A料,受智能終端設備巨大市場潛力的驅動,PPGi檢測生理信息的應用前景廣大,它對移動醫(yī)療的推廣和普及具有重要的現(xiàn)實意義,可以稱之為泛在式生命體征檢測的一個新方法。然而由于PPGi受到光照環(huán)境、攝像頭對光的敏感度、低采樣率和運動偽跡等因素的限制,目前PPGi的應用僅僅局限于心率和呼吸率的檢測�；谝陨蠁栴},本文首先針對在不同光照環(huán)境下PPGi特征進行深入分析,然后開發(fā)基于PPGi的容積脈搏波特征點提取方法,最后提出基于PPGi的心率、脈搏波傳導時間(Pulse Transit Time,PTT)和血氧飽和度(Blood Oxygen Saturation,Sp O2)檢測方法。本文開展以下幾項研究:(1)研究在不同光照環(huán)境下PPGi的特征。光源波長和攝像頭對PPGi有著重要的影響,實驗分別使用不同波長光和光強作為照明光評估基于PPGi特征。研究結果表明:不同波長光下基于PPGi計算的心率與金標準(金標準指當前臨床醫(yī)學界公認的診斷疾病的最可靠、最準確、最好的診斷方法)心率具有良好的一致性,僅僅在470nm、490nm、505nm、585nm、600nm、610nm、625nm、660nm波長光和白光作為照明光下,基于PPGi的容積脈搏波與金標準PPG信號波形具有良好的相關性;在600nm、610nm、625nm、660nm波長光和白光作為照明光下,隨著光強增加,基于PPGi的容積脈搏波直流強度增加,但是交流強度有一個增加到下降的趨勢。這些結果表明照明光的波長不會影響到PPGi計算心率,而照明光的波長和光的強度會影響PPGi的特征,為深入挖掘基于PPGi的生理信息檢測提供理論基礎。(2)研究基于PPGi的容積脈搏波特征點提取方法。針對基于PPGi的容積脈搏波具有基線漂移較大、采樣率低及分光現(xiàn)象等問題,提出基于PPGi的容積脈搏波特征點提取方法:提出使用重排平滑偽Wigner-Ville分布方法計算基于PPGi的容積脈搏波交流和直流強度;提出“自適應二次曲線門限”方法檢測基于PPGi的容積脈搏波峰和波谷;提出“二次曲線擬合”方法提高基于PPGi的容積脈搏波峰和波谷檢測準確度;提出“自適應區(qū)域變化”方法消除PPGi分光現(xiàn)象。通過計算基于PPGi的多個體表部位(手指端和太陽穴)心率和PTT,驗證此方法可以提高基于PPGi的容積脈搏波特征點檢測精度。(3)研究基于PPGi的心率檢測方法。首先評估基于PPGi的計算心率作為人臉區(qū)域特征的可行性;然后基于此原理,從臉部提取PPGi作為獨特的特征,區(qū)分人臉顏色相似背景環(huán)境中的人臉輪廓。使用相鄰4像素方法加強人臉區(qū)域可靠性,使用羅伯特算子創(chuàng)建臉部形狀區(qū)域。實驗結果表明基于PPGi的人臉檢測方法可以提高復雜和具有相似背景場景下人臉檢測的魯棒性。(4)研究基于PPGi的Sp O2檢測方法。首先在已分析不同波長光下基于PPGi的容積脈搏波特征的基礎上,選擇660nm和800nm波長光作為照明光,同時提取人體右手食指端與右手大拇指端的PPGi;然后在已建立基于PPGi的容積脈搏波特征點提取方法的基礎上,獲得基于PPGi的容積脈搏波特征點的交流和直流強度,并計算SpO2。實驗結果表明,與金標準Sp O2比較,95%的置信區(qū)間分布范圍達到?%2,體現(xiàn)良好的一致性。
[Abstract]:The pulse wave contains the abundant life information. Based on the pulse wave shape, intensity, speed and rhythm analysis, can get the physiological and pathological characteristics. Using the camera under ambient light conditions can extract pulse wave, it is through the use of volume pulse imaging (Photoplethysmographic imaging PPGi) method for detection of blood volume changes, through calculation PPGi image pixel average the two-dimensional signal into a one-dimensional signal can obtain volume pulse wave. The growing popularity of smart mobile phone with a camera and personal mobile terminal equipment to provide a low cost and ubiquitous health perception platform, people can not in the hospital, to evaluate the personal health status using the real-time dynamic mobile terminal equipment long time the non restraint environment, to achieve early warning of the disease, early prevention, early detection, early diagnosis and treatment purposes. It can be expected that by Driving intelligent terminal equipment huge market potential and application prospect of PPGi detection of physiological information is vast, it has important practical significance for the promotion and popularization of mobile medical, can be called the pan in a new method of vital signs detection. However due to the PPGi by the light environment, the sensitivity of the low light camera. The sampling rate and motion artifacts and other factors, the current PPGi application only to detect the heart rate and respiratory rate. Based on the above problems, this paper firstly according to the PPGi characteristics in different light environment in-depth analysis, and then develops a method for extracting feature points of pulse wave based on PPGi, and finally puts forward the PPGi based on the heart rate. Pulse transit time (Pulse Transit, Time, PTT) and oxygen saturation (Blood Oxygen Saturation, Sp O2) detection method. This paper carried out the following research: (1) study in different light environments PPGi The wavelength of light source and camera features. Has an important influence on the PPGi, experiments were made using different wavelengths of light intensity and light for evaluation based on the characteristics of PPGi. The results show that the PPGi calculation of the heart rate and the gold standard based on different wavelengths of light (the gold standard to diagnose diseases of current clinical medical profession recognized the most reliable, most accurate the best diagnostic method, heart rate) has good consistency, only in 470nm, 490nm, 505nm, 585nm, 600nm, 610nm, 625nm, 660nm wavelength and white light for PPGi, the volume pulse wave with the gold standard PPG signal waveform has a good correlation based in 600nm, 610nm, 625nm; 660nm. Wavelength and white light for, with the increase of light intensity, the increase of volume pulse wave DC intensity based on PPGi, but the exchange strength there is an increase to a downward trend. These results show that the wavelength of the illumination light does not affect PPGi to calculating the heart rate, and the light intensity of the light wavelength and light will affect the characteristics of PPGi, and provide a theoretical basis for further mining the physiological information detection based on PPGi. (2) research on pulse wave feature extraction method based on PPGi. Based on the PPGi volume pulse wave with base line drift is large, low sampling rate and the problem of optical phenomena, extraction method of volume pulse wave feature points based on PPGi put forward the pseudo Wigner-Ville distribution calculation for PPGi volume pulse wave AC and DC strength based on the use of rearrangement smoothing; put forward "two curve adaptive detection threshold method PPGi volume pulse peak and trough based on the proposed" two curve fitting; "the way to improve the PPGi volume pulse peak and trough detection based on accuracy; put forward the" adaptive regional change "method to eliminate PPGi light phenomenon. By calculating the individual based on PPGi Table parts (finger and temple) and heart rate PTT, verify that this method can improve the volume pulse wave feature point detection accuracy based on PPGi. (3) research on heart rate detection method based on PPGi. Firstly, based on PPGi evaluation and calculation of heart rate as a regional feasibility of facial features; then based on the principle of extracting PPGi from the face as a unique the characteristics of face contour distinguish face similar color background. Using 4 adjacent pixels method to strengthen the face region reliability, using the Robert operator to create the shape of the face region. The experimental results show that the face detection method based on PPGi can improve the complexity and robustness of face detection with similar background scenes. (4) Sp O2 detection method based on PPGi. Firstly, based on the characteristics of pulse wave on PPGi has been analyzed under different wavelengths of light, 660nm and 800nm wavelength as the light, at the same time Extraction of human finger end with the thumb of the right hand side of the PPGi; and has been established in the volume pulse wave extraction method based on feature points based on PPGi, obtain the volume pulse wave characteristics of PPGi AC and DC based on strength calculation and experimental results show that the SpO2. Sp O2 and the gold standard comparison, 95% confidence interval the distribution range of up to?%2, reflect the good consistency.

【學位授予單位】：哈爾濱工業(yè)大學
【學位級別】：博士
【學位授予年份】：2016
【分類號】：TN911.23;R318

【相似文獻】

相關期刊論文 前1條

1 崔波;高博;龔敏;;基于紅外反射光容積脈搏波的血管阻力研究[J];光散射學報;2011年04期

相關會議論文 前2條

1 趙蘭平;楊學智;牛欣;;指端容積脈搏波的動態(tài)血氧飽和度測定及應用[A];中國生理學會第23屆全國會員代表大會暨生理學學術大會論文摘要文集[C];2010年

2 宋娟;張松;羅志昌;楊益民;;應用指端容積脈搏波檢測與預測妊高征及其機理研究[A];中國生物醫(yī)學工程學會第六次會員代表大會暨學術會議論文摘要匯編[C];2004年

相關博士學位論文 前2條

1 劉赫;容積脈搏成像特征提取方法研究及在生理信號檢測的應用[D];哈爾濱工業(yè)大學;2016年

2 丁海泉;無創(chuàng)血糖檢測中的近紅外血流容積光譜基本問題研究[D];中國科學院研究生院（長春光學精密機械與物理研究所）;2010年

相關碩士學位論文 前3條

1 張雪;基于容積脈搏波的微循環(huán)特性研究[D];燕山大學;2012年

2 丁有得;基于容積脈搏波血流多參數(shù)測量的研究[D];南方醫(yī)科大學;2010年

3 王錦;電刺激小腦頂核對容積脈搏波參數(shù)影響的研究[D];重慶大學;2010年

，

本文編號：1363395