發(fā)布時間：2018-05-29 13:11

  本文選題：數(shù)字散斑相關(guān) + 位移與應(yīng)變測量��； 參考：《北京林業(yè)大學(xué)》2014年博士論文

【摘要】：數(shù)字散斑相關(guān)方法作為一種全場位移和應(yīng)變測量方法,在理論方面得到不斷發(fā)展和完善的同時,在工程應(yīng)用方面也開展了大量的工作。但目前的方法在求解效率和測量精度以及使用的通用性方面還存在一些不足。為了更好地滿足工程測試要求,本文對現(xiàn)有的數(shù)字散斑相關(guān)方法進行了進一步的研究與改進,論文的主要研究內(nèi)容和創(chuàng)新點如下。 本文針對現(xiàn)有的整像素逐點搜索算法對數(shù)字散斑相關(guān)方法搜索效率的影響,提出了一種基于連續(xù)性變形假設(shè)的鄰近域搜索算法。該算法首先通過逐點搜索法確定參考子區(qū)Ω在變形后圖像中的目標(biāo)子區(qū)Q’的位置后,根據(jù)連續(xù)性變形假設(shè),只需要在目標(biāo)子區(qū)Ω’附近的一個較小區(qū)域搜索即可,故只需要對第一個參考子區(qū)進行全場逐點搜索,其余子區(qū)都可以小范圍搜索,結(jié)果可大幅提高搜索效率。針對傳統(tǒng)的基于灰度梯度的亞像素位移求解算法中采用的灰度不變假設(shè)對位移求解精度的影響,將灰度線性變化引入基于梯度的亞像素位移算法。該算法假設(shè)變形前后圖像內(nèi)同一點的灰度由于光強波動而發(fā)生線性變化,引入灰度變化系數(shù)λ,應(yīng)用最小二乘迭代得到亞像素位移,解決了傳統(tǒng)梯度法由于光強變化引起的測量誤差問題。同時,定性地從圖像噪聲、亞像素插值誤差和形函數(shù)誤差三個方面分析了基于梯度的數(shù)字散斑相關(guān)測量方法的測量誤差,得到位移誤差標(biāo)準(zhǔn)差的數(shù)學(xué)表達式。 在影響數(shù)字散斑相關(guān)位移求解精度的因素中,除了算法誤差外,還包括散斑圖質(zhì)量、鏡頭光學(xué)失真等外界因素。為了提高散斑圖質(zhì)量進而提高數(shù)字散斑相關(guān)測量精度,本文提出一種新的模擬二值散斑圖,其可以在保證散斑顆粒隨機分布的同時通過判別散斑顆粒的位置與尺寸,避免出現(xiàn)傳統(tǒng)模擬散斑圖散斑顆粒過稀或過密的情況。該方法為模擬散斑圖的實現(xiàn)提供了一條新思路。在人工散斑制作方面,本文將碳粉熱轉(zhuǎn)印技術(shù)引入散斑制作,該方法通過熱壓原理將設(shè)計打印好的隨機散斑圖樣轉(zhuǎn)印在試件表面,具有能夠控制散斑顆粒的大小及分布密度和可重復(fù)性等優(yōu)點,避免了傳統(tǒng)噴涂散斑操作中由于人工經(jīng)驗不足造成的誤差,為人工散斑的制作提供了一條新的途徑。針對鏡頭光學(xué)失真對數(shù)字散斑相關(guān)測量精度的影響,本文在已有鏡頭畸變參數(shù)估計算法的基礎(chǔ)上,采用Harris法對標(biāo)定模板的特征點進行亞像素提取,并同時考慮鏡頭的徑向與切向畸變,對數(shù)字散斑相關(guān)測量試驗中所使用的CMOS鏡頭進行了非線性畸變系數(shù)估計與校正。通過試驗驗證了鏡頭畸變校正對數(shù)字散斑相關(guān)位移測量的重要性。 在數(shù)字散斑相關(guān)方法的應(yīng)變求解中,傳統(tǒng)的位移微分過程會放大位移場中所包含的噪聲。為了提高應(yīng)變求解精度,本文將有限元思想引入數(shù)字散斑相關(guān)應(yīng)變場的計算,提出了一種基于四邊形單元的應(yīng)變計算方法。首先,利用數(shù)字散斑相關(guān)方法獲得的位移場建立4節(jié)點和9節(jié)點四邊形單元,通過對應(yīng)的狀態(tài)方程計算出對應(yīng)節(jié)點的位移值,然后反插值計算出對應(yīng)單元內(nèi)每個點的應(yīng)變進而求得應(yīng)變場。最后,通過無疵和含孔試件的拉伸試驗,驗證了本文提出的算法能夠有效地提高應(yīng)變計算精度。 為了證明本文提出的改進數(shù)字散斑相關(guān)方法在工程應(yīng)用中的有效性和適應(yīng)性,以三個典型的工程測試問題為例進行了討論。應(yīng)用實例一是與溫控試驗箱結(jié)合建立了非接觸變溫變形測量系統(tǒng),用該系統(tǒng)測量了45號鋼在20℃～200℃溫度范圍內(nèi)的全場熱變形和熱膨脹系數(shù)。應(yīng)用實例二是在斷裂力學(xué)中的應(yīng)用,通過木材單邊裂紋拉伸試驗,得到了落葉松的復(fù)合型應(yīng)力強度因子和裂尖塑性區(qū)尺寸。應(yīng)用實例三是在蠕變力學(xué)中的應(yīng)用,結(jié)合自行設(shè)計的機械式蠕變試驗機對45號鋼的室溫蠕變進行測量,得到了不同應(yīng)力水平下的全場應(yīng)變和蠕變應(yīng)變-時間曲線。以上實例驗證了本文所提出的改進的數(shù)字散斑相關(guān)方法應(yīng)用于工程測試的可行性,表明了該方法的實用性和適應(yīng)性。
[Abstract]:As a method of full field displacement and strain measurement, the digital speckle correlation method has been developed and perfected in theory, and a great deal of work has been carried out in engineering application. However, there are still some shortcomings in the current method in solving the efficiency, measurement accuracy and the universality of the use. In this paper, the existing digital speckle correlation method is further studied and improved. The main research contents and innovations are as follows.
In this paper, a neighborhood search algorithm based on continuous deformation hypothesis is proposed to determine the effect of the existing integer pixel search algorithm on the search efficiency of the digital speckle correlation method. First, the algorithm determines the position of the target subregion Q 'of the reference subregion Omega in the deformed image by the point by point search method. We only need to search for a small area near the target region Omega, so it is only necessary to search the first reference subregion by the whole field, and the remaining subregions can be searched in a small range. The results can greatly improve the search efficiency. The gray invariant hypothesis used in the traditional gray gradient based subimage displacement algorithm is used. The gray linear change is introduced to the gradient based subpixel displacement algorithm. The algorithm assumes that the gray level of the same point in the image before and after the deformation is linearly changed due to the intensity fluctuation. The gray change coefficient is introduced into the gray level, and the subpixel displacement is obtained by using the least square iteration. The traditional gradient method is used to solve the light intensity. At the same time, the measurement error of the gradient based digital speckle correlation measurement method is analyzed qualitatively from three aspects of image noise, sub-pixel interpolation error and shape function error, and the mathematical expression of the standard deviation of displacement error is obtained.
In order to improve the quality of speckle pattern and the optical distortion of the lens, in addition to the algorithm error, in order to improve the speckle pattern quality and improve the accuracy of the digital speckle correlation measurement, a new simulated two value speckle pattern is proposed in this paper, which can ensure the random distribution of speckle particles. At the same time, by judging the position and size of speckle particles, the traditional simulated speckle pattern speckle particles are not too thin or overdense. This method provides a new idea for the realization of the simulation of speckle patterns. In the field of artificial speckle making, this paper introduces the thermal transfer technology of carbon powder into speckle making, and the method will be designed by the principle of hot pressing. The printed random speckle pattern is printed on the surface of the specimen, which can control the size, distribution density and repeatability of the speckle particles. It avoids the error caused by the shortage of artificial experience in the traditional spray speckle operation, and provides a new way for the fabrication of artificial speckle. On the basis of the estimation algorithm of the lens distortion parameters, the Harris method is used to extract the feature points of the calibration template, and the radial and tangential distortion of the lens are considered at the same time. The nonlinear distortion coefficient of the CMOS lens used in the digital speckle correlation measurement test is estimated and corrected. The importance of lens distortion correction to digital speckle correlation displacement measurement is verified by experiments.
In solving the strain solution of the digital speckle correlation method, the traditional displacement differential process magnifies the noise contained in the displacement field. In order to improve the accuracy of the strain solution, this paper introduces the finite element method into the calculation of the digital speckle correlation strain field, and proposes a strain calculation method based on the quadrilateral element. First, the digital speckle phase is used. The displacement field obtained by the closed method is established by the 4 node and the 9 node quadrilateral element, and the corresponding displacement value is calculated by the corresponding state equation. Then the strain of each point in the corresponding unit is calculated by the inverse interpolation and then the strain field is obtained. Finally, it is proved that the algorithm proposed in this paper can be effective through the tensile test of the free and porous specimen. The accuracy of the strain calculation is improved.
In order to prove the effectiveness and adaptability of the improved digital speckle correlation method in engineering application, three typical engineering testing problems are discussed. An application example is used to establish a non contact variable temperature deformation measurement system combined with temperature control test box. The system has been used to measure the temperature range of No. 45 steel at 20 C to 200 c. The full field thermal deformation and thermal expansion coefficient in the surrounding area. Application example two is applied in fracture mechanics. Through the tensile test of single side crack of wood, the composite stress intensity factor and crack tip plastic zone size of larch are obtained. Application example three is applied in Creep Mechanics, and the mechanical creep test machine designed by itself is applied to No. 45 The room temperature creep of steel is measured and the full field strain and creep strain time curve under different stress levels are obtained. The above examples verify the feasibility of the improved digital speckle correlation method proposed in this paper for engineering testing, which shows the practicability and adaptability of the method.
【學(xué)位授予單位】：北京林業(yè)大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2014
【分類號】：TP391.41;TB22

【參考文獻】

相關(guān)期刊論文 前10條

1 王顯;馬少鵬;陳俊達;宋義敏;;數(shù)字散斑相關(guān)方法的全場優(yōu)化表述[J];北京理工大學(xué)學(xué)報;2011年05期

2 伍曉贊;徐慧;王德志;;單邊斜裂紋的應(yīng)力強度因子計算[J];湖南文理學(xué)院學(xué)報(自然科學(xué)版);2009年01期

3 孫偉;何小元;胥明;羅斌;;數(shù)字圖像相關(guān)方法在膜材拉伸試驗中的應(yīng)用[J];工程力學(xué);2007年02期

4 張蕊;賀玲鳳;胡斌;;數(shù)字圖像相關(guān)測量鋼絞線彈性模量的應(yīng)用研究[J];工程力學(xué);2011年09期

5 潘兵;謝惠民;;基于差分進化的數(shù)字圖像相關(guān)方法[J];光電子.激光;2007年01期

6 鄭翔;張明;楊�？�;焦聯(lián)聯(lián);何小元;;數(shù)字圖像相關(guān)法在稀土摻雜水凝膠電響應(yīng)行為測量中的應(yīng)用[J];高分子材料科學(xué)與工程;2010年09期

7 ;DIGITAL SPECKLE CORRELATION METHOD IMPROVED BY GENETIC ALGORITHM[J];Acta Mechanica Solida Sinica;2003年04期

8 侯振德,秦玉文;基于圖像分形相關(guān)位移測量新方法的研究[J];光學(xué)學(xué)報;2002年02期

9 潘兵;謝惠民;夏勇;王強;;數(shù)字圖像相關(guān)中基于可靠變形初值估計的大變形測量[J];光學(xué)學(xué)報;2009年02期

10 李善祥,孫一翎,李景鎮(zhèn);數(shù)字散斑相關(guān)測量中亞象素位移搜索的曲面擬合研究[J];光子學(xué)報;1999年07期

，

本文編號：1950996