【摘要】： 圖像復(fù)原是數(shù)字圖像處理領(lǐng)域中最重要、最基本的研究課題之一,具有重要的理論價(jià)值和實(shí)際意義。其目的就是要盡量的恢復(fù)被退化圖像的本來面目。 傳統(tǒng)的圖像復(fù)原方法,如逆濾波法、維納濾波法、卡爾曼濾波法、奇異值分解偽逆法、最大熵復(fù)原法等,或面臨著高維方程的計(jì)算問題,或要求恢復(fù)過程滿足廣義平穩(wěn)過程的假設(shè),這些均是使得圖像復(fù)原問題沒有廣泛應(yīng)用的根本原因。神經(jīng)網(wǎng)絡(luò)由于其固有的自學(xué)習(xí)、自適應(yīng)性、強(qiáng)魯棒性及并行處理方面的潛能,因此被用于解決圖像處理領(lǐng)域內(nèi)多種問題,用神經(jīng)網(wǎng)絡(luò)進(jìn)行退化圖像的復(fù)原便是其中的應(yīng)用之一。 本文在研究了基于Hopfield神經(jīng)網(wǎng)絡(luò)的圖像復(fù)原算法的基礎(chǔ)上,為克服Hopfield神經(jīng)網(wǎng)絡(luò)易于陷入局部極小值的缺點(diǎn),進(jìn)一步提高復(fù)原圖像的信噪比和視覺效果,研究了一種在Hopfield神經(jīng)網(wǎng)絡(luò)模型中引入暫態(tài)混沌機(jī)制和小波理論用于圖像復(fù)原算法的新方法,實(shí)驗(yàn)證明了改進(jìn)方法的有效性。 本文的主要工作包括: 1.討論了一種基于Hopfield神經(jīng)網(wǎng)絡(luò)的圖像復(fù)原算法。在分析網(wǎng)絡(luò)更新規(guī)則過程中,為降低時(shí)間和空間復(fù)雜度,將原神經(jīng)網(wǎng)絡(luò)復(fù)原算法中用每一個(gè)像元恢復(fù)電平值由對(duì)應(yīng)的M個(gè)神經(jīng)元狀態(tài)之和表示改進(jìn)為用神經(jīng)元狀態(tài)變量群加權(quán)方案來表示。使得在保證網(wǎng)絡(luò)良好容錯(cuò)性的同時(shí),減少神經(jīng)網(wǎng)絡(luò)的整體規(guī)模;為保證網(wǎng)絡(luò)能夠精確地收斂到全局最小,采用狀態(tài)連續(xù)變化的狀態(tài)變量替代原階躍取值的狀態(tài)變量。 2.討論了一種基于混沌Hopfield神經(jīng)網(wǎng)絡(luò)的圖像復(fù)原算法。為改善Hopfield神經(jīng)網(wǎng)絡(luò)易于陷入局部極小值的缺點(diǎn),在Hopfield神經(jīng)網(wǎng)絡(luò)模型中引入混沌機(jī)制,可以獲得比Hopfield神經(jīng)網(wǎng)絡(luò)更加豐富和更加靈活的動(dòng)力學(xué)特性,從而具有更強(qiáng)的搜索全局最優(yōu)解或近似全局最優(yōu)解的能力,較大程度提高了收斂性能和初值魯棒性。 3.討論了一種基于小波混沌神經(jīng)網(wǎng)絡(luò)的圖像復(fù)原算法。原網(wǎng)絡(luò)模型的激勵(lì)函數(shù)采用單調(diào)遞增的Sigmoid函數(shù),其逼近函數(shù)的能力沒有基函數(shù)強(qiáng),并且Sigmoid函數(shù)在逼近過程中會(huì)產(chǎn)生冗余。因此將小波理論引入混沌Hopfield神經(jīng)網(wǎng)絡(luò)中,構(gòu)造由小波函數(shù)和Sigmoid函數(shù)組成的新的激勵(lì)函數(shù),使得網(wǎng)絡(luò)具有更強(qiáng)的函數(shù)逼近能力。
[Abstract]:Image restoration is one of the most important and basic research topics in the field of digital image processing, which has important theoretical and practical significance. The aim is to restore the degraded image as much as possible. Traditional image restoration methods, such as inverse filtering, Wiener filtering, Kalman filtering, singular value decomposition pseudo-inverse, maximum entropy restoration, etc. The restoration process is required to satisfy the assumption of generalized stationary process, which is the fundamental reason that the image restoration problem is not widely used. Because of its inherent self-learning, self-adaptability, strong robustness and potential in parallel processing, neural networks are used to solve many problems in the field of image processing, and the restoration of degraded images using neural networks is one of the applications. In this paper, based on the research of image restoration algorithm based on Hopfield neural network, in order to overcome the shortcoming that Hopfield neural network is easy to fall into local minimum, and further improve the SNR and visual effect of reconstructed image. A new method for image restoration based on transient chaotic mechanism and wavelet theory in Hopfield neural network model is studied. The experimental results show that the improved method is effective. The main work of this paper includes: 1. An image restoration algorithm based on Hopfield neural network is discussed. In order to reduce the time and space complexity, The restoration level of each pixel in the original neural network restoration algorithm is improved from the sum representation of the corresponding M neuron states to the weighted scheme of neuron state variables group. In order to ensure that the network can converge to the global minimum accurately, the overall scale of the neural network can be reduced while ensuring the good fault-tolerance of the network. The state variable with continuous state change is used to replace the state variable of the original step value. 2. An image restoration algorithm based on chaotic Hopfield neural network is discussed. In order to improve the disadvantage that Hopfield neural network is prone to fall into local minima, chaotic mechanism can be introduced into the Hopfield neural network model to obtain more abundant and flexible dynamic characteristics than Hopfield neural network. Therefore, it has stronger ability to search global optimal solution or approximate global optimal solution, and improves convergence performance and initial value robustness to a great extent. An image restoration algorithm based on wavelet chaotic neural network is discussed. The excitation function of the original network model adopts the monotone increasing Sigmoid function, and its ability to approximate the function is not as strong as the basis function, and the Sigmoid function will produce redundancy in the process of approximation. Therefore, the wavelet theory is introduced into chaotic Hopfield neural network, and a new excitation function is constructed, which is composed of wavelet function and Sigmoid function, which makes the network have stronger function approximation ability.
【學(xué)位授予單位】：江蘇科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2010
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 嚴(yán)圣華;羅兵;;基于小波神經(jīng)網(wǎng)絡(luò)的視頻圖像恢復(fù)[J];長江大學(xué)學(xué)報(bào)(自科版)理工卷;2006年01期

2 何振亞,譚營,王保云;一種新的混沌神經(jīng)網(wǎng)絡(luò)模型及其動(dòng)力學(xué)分析[J];東南大學(xué)學(xué)報(bào);1998年06期

3 王學(xué)武,顧幸生;神經(jīng)網(wǎng)絡(luò)的研究熱點(diǎn)分析[J];工業(yè)儀表與自動(dòng)化裝置;2004年06期

4 張軍英,盧志軍,石林,董繼揚(yáng),石美紅;基于脈沖耦合神經(jīng)網(wǎng)絡(luò)的椒鹽噪聲圖像濾波[J];中國科學(xué)E輯:信息科學(xué);2004年08期

5 王凌,鄭大鐘;TSP及其基于Hopfield網(wǎng)絡(luò)優(yōu)化的研究[J];控制與決策;1999年06期

6 王磊,戚飛虎,莫玉龍;精確復(fù)原退化圖象的連續(xù) Hopfield 網(wǎng)絡(luò)研究[J];上海交通大學(xué)學(xué)報(bào);1997年12期

7 李翠華,鄭南寧;構(gòu)造徑向基函數(shù)的一般方法及其在圖象處理中的應(yīng)用[J];數(shù)值計(jì)算與計(jì)算機(jī)應(yīng)用;2000年02期

8 韓玉兵,吳樂南;基于狀態(tài)連續(xù)變化的Hopfield神經(jīng)網(wǎng)絡(luò)的圖像復(fù)原[J];信號(hào)處理;2004年05期

9 韓玉兵,吳樂南;基于改進(jìn)Paik型Hopfield網(wǎng)絡(luò)的圖像復(fù)原[J];應(yīng)用科學(xué)學(xué)報(bào);2005年02期

10 柏森,張邦禮,曹長修;神經(jīng)網(wǎng)絡(luò)圖象復(fù)原方法的研究進(jìn)展[J];中國圖象圖形學(xué)報(bào);2002年11期

相關(guān)博士學(xué)位論文 前1條

1 吳亞東;圖像復(fù)原算法研究[D];電子科技大學(xué);2006年

相關(guān)碩士學(xué)位論文 前2條

1 王輝;基于神經(jīng)網(wǎng)絡(luò)的退化圖像復(fù)原算法研究[D];上海交通大學(xué);2007年

2 岳海燕;混沌神經(jīng)網(wǎng)絡(luò)模擬退火參數(shù)研究[D];哈爾濱工程大學(xué);2007年

，

本文編號(hào)：2200072