CAJ | 학술논문

针对经典的总变分去噪模型边缘信息对噪声敏感且易模糊的缺陷,提出了非线性与线性的加权变分模型.非线性加权变分模型是在总变分模型的正则项中引入权函数,并利用权函数引导扩散,使得新模型在消噪的同时更好地保持图像的纹理特征和边缘信息;线性加权变分模型是对含噪图利用高斯函数进行预处理,再对处理后的图像进行扩散,从而降低计算复杂度.数值实验表明,与经典的总变分模型相比,改进的方法无论是在视觉效果还是峰值信噪比上都有明显的提高.
침대경전적총변분거조모형변연신식대조성민감차역모호적결함,제출료비선성여선성적가권변분모형.비선성가권변분모형시재총변분모형적정칙항중인입권함수,병이용권함수인도확산,사득신모형재소조적동시경호지보지도상적문리특정화변연신식;선성가권변분모형시대함조도이용고사함수진행예처리,재대처리후적도상진행확산,종이강저계산복잡도.수치실험표명,여경전적총변분모형상비,개진적방법무론시재시각효과환시봉치신조비상도유명현적제고.
View on the weakness of a classical variational de-noising model in which edge information is sensitive to noises and prone to blur, two improved nonlinear and linear weighted variational algorithms are put forward. In the nonlinear weighted variational model, a weight function is introduced in the regularization term of the classical model to induct diffusion, which gives the result that the new model preserves the texture characteristics and the edge information better while removing noises. In the linear model, the Gaussian function is used to smooth the noised image before diffusion, which reduces the computational complexity. Compared with the classical model, the experimental results show improvements of both the proposed models.