【摘要】 图像的复扩散是图像增强的一种有效方法.传统的复扩散算法在求解过程中利用迭代产生的虚部,即图像高斯卷积的拉普拉斯算子作为扩散控制因子.然而由于该因子的平滑、时间依赖以及二阶特性,图像在处理后会在边缘处形成带状模糊.为此,本文提出一种改进的结合了图像迭代实边缘特性的复扩散算法.实验结果显示,该算法可以更好的增强模糊图像的边缘和去除噪声.更多还原

【Abstract】 Complex diffusion is one of the effective methods for image enhancement.The conventional complex diffusion algorithm uses the iterative imaginary part(Laplacian of the Gaussian kernel) to control the diffusion process.But for some attributes of the imaginary part,such as smoothness,time-dependent and second order derivative,there would be belt-like blurred area near the edges.Aiming at this drawback,this paper developed a new complex diffusion algorithm based on iterative real edge attributes.The results show that compared with the conventional algorithms,the new one can achieve better enhancing effects for blurred and noisy images.更多还原