天津科技大学学报
天津科技大學學報
천진과기대학학보
JOURNAL OF TIANJIN UNIVERSITY OF SCIENCE & TECHNOLOGY
2014年
6期
74-77
,共4页
正则化%反问题%图像重建
正則化%反問題%圖像重建
정칙화%반문제%도상중건
regularization%inverse-problem%image reconstruction
针对电学层析成像技术的反问题求解,提出一种新的正则化图像重建算法及正则参数选择方法。采用最小二乘拟合法构建正则参数与评价参数的 Matlab 数学模型,取极值获得正则参数值;利用 COMSOL Multiphysics 建立仿真场域,最后利用 Matlab 求解反问题,重建图像。通过对重建图像及评价参数的对比分析,验证了方法的有效性。结果表明,所用方法及正则参数能够有效改善反问题的病态性,使反问题的求解质量最优。
針對電學層析成像技術的反問題求解,提齣一種新的正則化圖像重建算法及正則參數選擇方法。採用最小二乘擬閤法構建正則參數與評價參數的 Matlab 數學模型,取極值穫得正則參數值;利用 COMSOL Multiphysics 建立倣真場域,最後利用 Matlab 求解反問題,重建圖像。通過對重建圖像及評價參數的對比分析,驗證瞭方法的有效性。結果錶明,所用方法及正則參數能夠有效改善反問題的病態性,使反問題的求解質量最優。
침대전학층석성상기술적반문제구해,제출일충신적정칙화도상중건산법급정칙삼수선택방법。채용최소이승의합법구건정칙삼수여평개삼수적 Matlab 수학모형,취겁치획득정칙삼수치;이용 COMSOL Multiphysics 건립방진장역,최후이용 Matlab 구해반문제,중건도상。통과대중건도상급평개삼수적대비분석,험증료방법적유효성。결과표명,소용방법급정칙삼수능구유효개선반문제적병태성,사반문제적구해질량최우。
Focused on solving the inverse-problem of electrical tomography,a new regularization image reconstruction algorithm and a method to select the regularization parameters were proposed. A Matlab mathematic model of regularization parameter and evaluation parameter was constructed by the least square fit method.The values of regularization parameter can be obtained by taking the extreme.COMSOL Multiphysics was adopted to construct the simulation field,and Matlab was used to solve the inverse-problem and reconstruct the image. The effectiveness of the method was verified by contrasting the reconstructed image and the evaluation parameters. The result indicates that the method and regularization parameters can improve the ill-pose of the inverse-problem effectively and make the solution to the inverse problem optimal.