基于张量投票的图像去噪算法研究

针对传统的基于张量投票算法图像去噪过程中对于对称图像存在的张量叠加计算量冗余的问题,提出了张量投票域的改进算法。首先,利用基于相位信息的图像对称性检测方法对需要去噪的图像进行检测。然后,根据检测出的图像的对称性的特性简化投票域的区域。最后,将改进后的投票域应用在图像去噪过程中。实验仿真的结果表明,这里的方法不仅明显提高了图像的峰值信噪比,而且较好地保护了图像的细节信息,在算法的效率方面也有较明显的提高。

An Image Denoising Algorithm based on Tensor Voting

For the redundant tensor voting superposition comptutation existing in the traditional tensor voting of symmetrical image denoising,an improved algorithm of tensor voting field is proposed.First,based on phase information,the symmetry property of the image is detected.Then the voting field is simplified according to the detected result.Finally this simplified tensor voting field is employed in the process of image denoising.The simulation result shows that the proposed method could efficiently reduce the im...