磁共振力顯微鏡的信號處理與圖像復原研究
發(fā)布時間：2018-09-12 05:16
【摘要】：磁共振力顯微鏡是一種能實現(xiàn)納米級分辨率的三維分子成像儀器,它結合了磁共振成像技術與掃描探針顯微術,具有磁共振成像的三維探測能力與掃描探針顯微術的納米分辨能力,可以實現(xiàn)單個核自旋的探測。然而,磁共振力顯微鏡在實際應用中存在掃描速度慢、易受噪聲影響和成像速度慢等問題,解決這些問題有賴于有效、精妙的信號處理方法和圖像處理方法的運用。本文研究了磁共振力顯微鏡涉及的一些的信號處理方法和圖像處理方法,包括四個方面:帶噪正弦信號的參數(shù)估計、ARMA模型在微懸臂的系統(tǒng)辨識中的應用、樣品自旋密度圖像的復原、擴頻及解擴方案的設計。主要的研究工作和創(chuàng)新點如下:(1)本文研究了帶噪正弦信號的參數(shù)估計。根據(jù)信號攜帶的噪聲的統(tǒng)計特性,使用分段幅度及相位聯(lián)合最大似然估計來估計帶噪正弦信號的幅度及相位,仿真研究表明估計算法的估計誤差的方差很接近Cramer-Rao下界,估計性能良好。(2)為了了解微懸臂的模態(tài)信息,本文采用ARMA模型對微懸臂進行系統(tǒng)辨識和模態(tài)參數(shù)識別。針對觀測噪聲會使時間序列模型的辨識精度變低的問題,本文研究了如何將帶有觀測噪聲的ARMA模型轉化為無觀測噪聲的ARMA模型。得到微懸臂的ARMA模型后將這個模型轉化為連續(xù)系統(tǒng)的傳遞函數(shù),再從傳遞函數(shù)中求出微懸臂的各模態(tài)參數(shù)。為了減小微懸臂對外力的延遲響應時間從而提高磁共振力顯微鏡的掃描速度,本文根據(jù)系統(tǒng)辨識得到的模型,構造一個復原濾波器,微懸臂的振動信號通過這個復原濾波器后得到作用力信號。仿真及實驗結果表明了本文所用的系統(tǒng)辨識、模態(tài)參數(shù)識別及信號復原方法的有效性。(3)為了從退化圖像中恢復出自旋密度圖像,本文研究了磁共振力顯微鏡的成像機理和自旋密度圖像的復原方法。根據(jù)磁共振力顯微鏡的成像機理,推導了點擴散函數(shù)的表達式。當點擴散函數(shù)容易求得時,考慮圖像復原問題的病態(tài)性,用基于稀疏性約束的非盲圖像復原算法進行圖像復原;當點擴散函數(shù)不容易求得時,用盲圖像復原算法進行圖像復原,同時復原出點擴散函數(shù)和自旋密度圖像。(4)為了提高磁共振力顯微鏡的抗干擾性,本文提出了一個擴頻及解擴方案,這個方案對磁共振力顯微鏡的輸入信號進行擴頻,并對輸出信號進行解擴。本文用信號處理的理論研究了這個方案的抗干擾性,并與未施加擴頻的方案進行比較,發(fā)現(xiàn)在一定條件下提出的方案的抗干擾性更佳。最后用仿真實驗來證明這個方案的有效性。
[Abstract]:Magnetic resonance force microscopy (MRFM) is a three-dimensional molecular imaging instrument with nanometer resolution. It combines magnetic resonance imaging (MRI) with scanning probe microscopy (SPM). It has the three-dimensional detection ability of magnetic resonance imaging (MRI) and the nanometer resolution ability of scanning probe microscopy (SPM). It can detect the spin of a single nucleus. There are some problems in practical application, such as slow scanning speed, easy to be affected by noise and slow imaging speed. To solve these problems, it is necessary to use effective and delicate signal processing methods and image processing methods. Parameter estimation of sinusoidal signal, application of ARMA model in system identification of micro-cantilever, restoration of sample spin density image, design of spread spectrum and despreading scheme. The main research work and innovation are as follows: (1) Parameter estimation of noisy sinusoidal signal is studied in this paper. Bit joint maximum likelihood estimation is used to estimate the amplitude and phase of the noisy sinusoidal signal. Simulation results show that the variance of the estimation error is close to Cramer-Rao lower bound and the estimation performance is good. (2) In order to understand the modal information of the micro-cantilever, the ARMA model is used to identify the system and modal parameters of the micro-cantilever. In this paper, we study how to transform the ARMA model with observation noise into the ARMA model without observation noise. After obtaining the ARMA model of the micro-cantilever, the model is transformed into the transfer function of the continuous system, and then the modal parameters of the micro-cantilever are obtained from the transfer function. In order to reduce the delay response time of the micro-cantilever to the external force and improve the scanning speed of the magnetic resonance force microscope, a restoring filter is constructed according to the model of the system identification. The vibration signal of the micro-cantilever passes through the restoring filter to get the force signal. (3) In order to recover the spin density image from the degraded image, the imaging mechanism of the magnetic resonance force microscope and the restoration method of the spin density image are studied in this paper. Then, considering the ill-conditioned of the image restoration problem, a non-blind image restoration algorithm based on sparse constraints is used to restore the image. When the point spread function is not easy to obtain, the blind image restoration algorithm is used to restore the image, and the point spread function and the spin density image are restored simultaneously. (4) To improve the anti-interference ability of the magnetic resonance force microscope, In this paper, a spread spectrum and despread scheme is proposed, which spread spectrum the input signal of the magnetic resonance force microscope and despread the output signal. The anti-jamming performance of this scheme is studied with the theory of signal processing, and compared with the scheme without spread spectrum, it is found that the anti-jamming of the scheme proposed under certain conditions. It is proved that the scheme is effective.
【學位授予單位】：中國科學院大學(中國科學院武漢物理與數(shù)學研究所)
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41
本文編號：2238019
[Abstract]:Magnetic resonance force microscopy (MRFM) is a three-dimensional molecular imaging instrument with nanometer resolution. It combines magnetic resonance imaging (MRI) with scanning probe microscopy (SPM). It has the three-dimensional detection ability of magnetic resonance imaging (MRI) and the nanometer resolution ability of scanning probe microscopy (SPM). It can detect the spin of a single nucleus. There are some problems in practical application, such as slow scanning speed, easy to be affected by noise and slow imaging speed. To solve these problems, it is necessary to use effective and delicate signal processing methods and image processing methods. Parameter estimation of sinusoidal signal, application of ARMA model in system identification of micro-cantilever, restoration of sample spin density image, design of spread spectrum and despreading scheme. The main research work and innovation are as follows: (1) Parameter estimation of noisy sinusoidal signal is studied in this paper. Bit joint maximum likelihood estimation is used to estimate the amplitude and phase of the noisy sinusoidal signal. Simulation results show that the variance of the estimation error is close to Cramer-Rao lower bound and the estimation performance is good. (2) In order to understand the modal information of the micro-cantilever, the ARMA model is used to identify the system and modal parameters of the micro-cantilever. In this paper, we study how to transform the ARMA model with observation noise into the ARMA model without observation noise. After obtaining the ARMA model of the micro-cantilever, the model is transformed into the transfer function of the continuous system, and then the modal parameters of the micro-cantilever are obtained from the transfer function. In order to reduce the delay response time of the micro-cantilever to the external force and improve the scanning speed of the magnetic resonance force microscope, a restoring filter is constructed according to the model of the system identification. The vibration signal of the micro-cantilever passes through the restoring filter to get the force signal. (3) In order to recover the spin density image from the degraded image, the imaging mechanism of the magnetic resonance force microscope and the restoration method of the spin density image are studied in this paper. Then, considering the ill-conditioned of the image restoration problem, a non-blind image restoration algorithm based on sparse constraints is used to restore the image. When the point spread function is not easy to obtain, the blind image restoration algorithm is used to restore the image, and the point spread function and the spin density image are restored simultaneously. (4) To improve the anti-interference ability of the magnetic resonance force microscope, In this paper, a spread spectrum and despread scheme is proposed, which spread spectrum the input signal of the magnetic resonance force microscope and despread the output signal. The anti-jamming performance of this scheme is studied with the theory of signal processing, and compared with the scheme without spread spectrum, it is found that the anti-jamming of the scheme proposed under certain conditions. It is proved that the scheme is effective.
【學位授予單位】：中國科學院大學(中國科學院武漢物理與數(shù)學研究所)
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41
【參考文獻】
相關期刊論文 前1條
1 曹更玉;付號;;磁共振力顯微鏡[J];電子顯微學報;2011年01期
相關博士學位論文 前2條
1 李鑫楠;圖像盲復原算法研究[D];吉林大學;2015年
2 黃麗麗;圖像復原中若干問題的正則化模型與算法[D];南京理工大學;2013年
相關碩士學位論文 前2條
1 莫平杰;基于HHT方法的機械系統(tǒng)模態(tài)參數(shù)識別[D];浙江大學;2011年
2 單曉明;數(shù)字圖像的盲復原研究[D];南京理工大學;2005年
，本文編號：2238019
本文鏈接：http://www.lk138.cn/shoufeilunwen/xixikjs/2238019.html
最近更新
教材專著
