請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15502
標題: | 基於小波函數解析之創新光學系統研究與開發:光子都卜勒干涉儀及生理訊號量測 Research and Development of Wavelet-based Innovative Optical Instruments: PDV and Physiological Signals Measurement |
作者: | Chun-Hsiung Wang 王俊雄 |
指導教授: | 李世光(Chi-Kung Lee) |
關鍵字: | 小波轉換,複合材料,速度量測,表面輪廓量測,生理訊號量測, wavelet,composite,velocity measurement,surface morphology,physiological signals, |
出版年 : | 2020 |
學位: | 博士 |
摘要: | 小波轉換因具備可調變的頻率解析度、快速的運算效率以及針對瞬時變化的頻率成分敏感度等特性而被廣泛運用。在本論文中,即利用小波轉換本身優異的特性,輔助數項創新型光學量測系統技術的開發,以單點速度量測出發到全域動態訊號量測,包含超高速速度量測儀、基於條紋解相技術的表面輪廓量測、以手腕橈動脈收縮行為觀測為主的生理訊號量測。 首先,本論文基於複合材料在高速碰撞之破壞行為研究之需求下,開發高速噴射碎片產生器及基於光學都卜勒干涉儀架構之超高速速度量測儀。噴射碎片產生器透過不同高電壓施以電橋所產生之電弧爆炸行為,提供在安全的操作情況下達成模擬不同速度範圍之撞擊測試。所開發的超高速度量測儀,以商用波段光纖降低成本,同時達到簡潔的架構、以及達每秒千公尺等級的速度量測需求,搭配小波演算法及自行開發的速度歷程曲線追跡演算法,已成功驗證並且量測數種動能驅動之複材飛行碎片之速度歷程,此噴射碎片產生器具備可更改口徑及可替換不同複合材料之多層結構,可提供各種複合材料在不同撞擊動能的破壞行為測試。 其二,非接觸式並可全域偵測物體高度的量測即時性已在各方面的應用上產生強烈需求,本論文即針對操作於頻譜域的條紋解相演算法進行完整審視,針對量測需求比較窗函數及小波函數差異,並將頻譜轉換因子之於相位計算的操作重新規劃。同時以模擬及實際架設實驗比較不同演算法之優劣特性,並且驗證所提出基於小波的條紋分析演算法,具備高準確性、高紋理還原度以及相對快的運算速度。所開發單張條紋解相演算法可應用於動態量測,並接續進行第三部分之橈動脈運動行為研究,以非接觸式量測方式對動脈進行動態量測。 第三方面,為能達到心血管疾病的監測與預防,新式血壓量測技術必須滿足四大條件:無需監督、不需配戴充氣袖帶、可連續量測、及高準確性,而近年所發展之無袖帶式血壓量測技術尚未滿足高準確性的要求。因此,本研究在陰影疊紋的架構下,利用第二部分所提出的條紋分析演算法在無受外力、非接觸的情形下,對於脈搏振動訊號進行量測,並且同時與其他生理訊號如心電圖、血流速波形、血流量波形、血壓值等參數做比較,量測結果清楚顯示手腕處橈動脈收縮運動行為與上述參數之相對關係,並可作為後續血壓模型發展作為基礎。 Wavelet transform is known for its benefits of versatile frequency localization accuracy, high computation efficiency, and high sensitivity to cope with signals with instantaneous frequency change. In this dissertation, we took the advantages of wavelet transform to perform signal analysis within the development of innovative optical metrology systems in three aspects: 1) ultra-high velocity measurement, 2) surface metrology with fringe analysis, and 3) physiological signals measurement. To provide the solution to the facture behavior evaluation under impact events of high momentum for composite materials, we developed a fragment generator design and a velocity measurement system based on photonic Doppler velocimetry. The continuous wavelet transform was adopted to generate the spectrum distribution over the time-frequency plane from the collected heterodyne signals, and we developed a velocity line tracing algorithm to acquire quantized velocity profile of the ejecting fragments introduced from the simulated impact events. The real-time requirement for contactless full-field spatial detection of an object’s surface has drawn much attention to meet various applications. We reviewed the fringe analysis algorithms operating at the spectrum domain, compared different transformation bases, and summarized the transformation product operations. The wavelet-based algorithm was shown with high accuracy, accurate reconstructed texture, and relatively fast computation speed. The proposed algorithm was used to analyze the shadow moiré fringe in the latter part of dynamic arterial vibration measurement. The development of blood pressure monitoring with unsupervised, cuff-less, continuous, and accurate methods has significant potential to control cardiovascular diseases (CVDs). The recent cuff-less methods still lack sufficient accuracy. We used the shadow moiré topography with our proposed fringe analysis algorithm to perform non-contact inspection on the arterial vibration. Compared with different physiological signals, the vessel contraction behavior at the radial artery was clearly classified and would provide strong base for further development of blood pressure models. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15502 |
DOI: | 10.6342/NTU202001496 |
全文授權: | 未授權 |
顯示於系所單位: | 應用力學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-1407202010154800.pdf 目前未授權公開取用 | 12.72 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。