本文關(guān)鍵詞： 微血管 圖像 自動(dòng)聚焦 稀疏去噪　出處：《東北石油大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：微血管參與有機(jī)體內(nèi)很多疾病的發(fā)生、發(fā)展過程,對(duì)于疾病的預(yù)警有很大的參考價(jià)值,并且提示了疾病臨床治療的新研究方向。要探討微血管在有機(jī)體內(nèi)的作用,首先要獲得清晰的微血管顯微圖像。但研究對(duì)象為活體狀態(tài),由于其生理活動(dòng),會(huì)導(dǎo)致圖像采集困難,并且圖像在采集過程中會(huì)因受到噪聲污染而使得成像質(zhì)量變差,為了達(dá)到采集清晰圖像的目的,設(shè)立了本課題,并進(jìn)行了以下研究:首先,設(shè)計(jì)并實(shí)現(xiàn)了微血管顯微圖像采集系統(tǒng)�；铙w小動(dòng)物自身的生理活動(dòng)會(huì)產(chǎn)生圖像待采集區(qū)域的平面位移或改變焦距,為了能采集到相對(duì)清晰的圖像,本研究應(yīng)用Visual Studio 2005在MFC的基礎(chǔ)上應(yīng)用C++程序設(shè)計(jì)語言編程實(shí)現(xiàn)了PC主機(jī)與二維運(yùn)動(dòng)平臺(tái)和工業(yè)攝像機(jī)的接口設(shè)計(jì),并應(yīng)用Visual Studio 2005在DirectX套件的基礎(chǔ)上應(yīng)用C++程序設(shè)計(jì)語言編程實(shí)現(xiàn)了PC主機(jī)與平臺(tái)運(yùn)動(dòng)控制器之間的接口設(shè)計(jì),保證平臺(tái)運(yùn)動(dòng)控制器對(duì)于二維運(yùn)動(dòng)平臺(tái)和工業(yè)攝像機(jī)的控制,實(shí)現(xiàn)了系統(tǒng)的圖像采集功能。然后,設(shè)計(jì)并實(shí)現(xiàn)了自動(dòng)調(diào)焦系統(tǒng)。在圖像采集系統(tǒng)的基礎(chǔ)上引進(jìn)一個(gè)步進(jìn)電機(jī),通過比較常用調(diào)焦函數(shù)特性,確定應(yīng)用Variance函數(shù)作為自動(dòng)調(diào)焦系統(tǒng)的運(yùn)算函數(shù),并基于逐次逼近法獲取了采集圖像的正焦位置,并在此基礎(chǔ)上設(shè)計(jì)了自動(dòng)調(diào)焦程序流程,應(yīng)用C++程序設(shè)計(jì)語言編程實(shí)現(xiàn)了圖像采集系統(tǒng)的自動(dòng)調(diào)焦功能。最后,對(duì)于獲取的微血管顯微圖像進(jìn)行稀疏去噪處理。圖像的稀疏去噪處理要解決稀疏去噪方法的選擇和字典的設(shè)計(jì)兩個(gè)關(guān)鍵性問題,通過對(duì)于常用算法收斂性和在構(gòu)造字典的過程中所用到的原子數(shù)的比較,本研究中選擇K-SVD算法訓(xùn)練出的字典對(duì)微血管顯微圖像進(jìn)行稀疏去噪處理。并基于這種理論進(jìn)行了仿真實(shí)驗(yàn),K-SVD訓(xùn)練出的學(xué)習(xí)字典的成像質(zhì)量好,本文提出的基于K-SVD訓(xùn)練出的學(xué)習(xí)字典具有良好的自適應(yīng)性。
[Abstract]:Microvasculature is involved in the occurrence and development of many diseases in organic body, which has great reference value for disease warning. In order to explore the role of microvasculature in the organism, we must obtain a clear micrograph of the microvessel, but the object of study is living state, because of its physiological activities. It will lead to the difficulty of image acquisition, and the image quality will become worse because of the noise pollution in the process of image acquisition. In order to achieve the purpose of collecting clear images, this subject has been set up. The following studies are carried out: firstly, the microvascular microimage acquisition system is designed and implemented. The physiological activities of the living small animals will result in the plane displacement of the image to be collected or the change of focal length. In order to capture relatively clear images. In this study, Visual Studio 2005 is applied to realize the interface design of PC host computer with two-dimensional motion platform and industrial camera based on MFC. And the application of Visual Studio 2005 in the DirectX suite based on the application of C. The interface design between PC host and platform motion controller is realized by programming language. Ensure that the platform motion controller for two-dimensional motion platform and industrial camera control, achieve the system image acquisition function. Then. An automatic focusing system is designed and implemented. Based on the image acquisition system, a step motor is introduced, and the characteristics of the common focusing function are compared. The Variance function is used as the operation function of the auto-focusing system, and the positive focus position of the collected image is obtained based on the successive approximation method, and the program flow of auto-focusing is designed on this basis. C programming language is used to realize the automatic focusing function of the image acquisition system. Finally. The sparse denoising of the microvascular microimages obtained is done. The two key problems of sparse denoising are the selection of sparse denoising methods and the design of dictionaries. The convergence of common algorithms and the number of atoms used in the construction of dictionaries are compared. In this study, the dictionary trained by K-SVD algorithm is selected to deal with sparse denoising of microvascular image, and the simulation experiment is carried out based on this theory. The learning dictionary trained by K-SVD has good imaging quality, and the learning dictionary based on K-SVD has good adaptability.
【學(xué)位授予單位】：東北石油大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：R318;TP391.41
，

本文編號(hào)：1454865