分数阶相位拉伸变换的图像锐化算法

IMAGE SHARPING ALGORITHM BASED ON FRACTIONAL PHASE STRETCH TRANSFORM

图像锐化是图像增强的主要内容之一,在图像分析、图像理解以及医学图像等领域均有重要的应用。现有图像锐化方法对图像中的弱强度变化特征增强效果不明显,并且在边缘附近还会出现毛刺与噪声。为解决这些问题,提出一种基于相位拉伸变换结合相对总变分的图像锐化算法,其主要工作包括分析推导相位拉伸变换中的核函数,将传统相位拉伸变换推广至分数阶傅里叶变换,然后与相对总变分理论相结合。实验结果表明,该算法对弱强度变化特征具有更好的锐化增强效果,同时由于相对总变分的约束,也消除了增强图像边缘附近的毛刺现象。与传统锐化增强算法相比,该算法的平均梯度和信息熵提高均在80%以上,证明了该算法的有效性与优越性。

Image sharpening is one of the main research topics in image enhancement and has important applications in image analysis,image understanding and medical imaging.The existing image sharpening methods do not have obvious enhancement effect on the weak intensity-change features in the image,and burr and noise appear near the edge.To solve these problems,we propose an image sharpening algorithm based on phase stretching transform and relative total variation in this paper.We analyzed and deduced the kernel function of the phase stretching transform.The traditional phase stretching transform was extended to the fractional Fourier transform,and the relative total variation was incorporated into image sharping.The experimental results show that the algorithm proposed in this paper has better sharpening and enhancement effect on weak intensity-change features,and due to the constraint of relative total variation,it eliminates burrs near the edge of the enhanced image.Compared with the traditional sharpening enhancement algorithm,the average gradient and information entropy of the algorithm in this paper are both increased by more than 80%,which proves the effectiveness and superiority of the proposed algorithm.