论文部分内容阅读
在受随机值脉冲噪声干扰的图像中,为了去除图像中的脉冲噪声并有效地保护图像的边缘与细节,提出了一种新的两阶段细节保护随机值脉冲噪声滤波算法.在噪声检测阶段,针对图像中边缘和细节像素难以和噪声像素有效区分的问题,提出了一种基于S-估计的绝对级差统计量(S-ROAD).通过引入S-估计到ROAD统计量,消除了ROAD数据中由图像边缘和细节带来的干扰.利用S-ROAD统计量,图像中的大部分噪声像素,包括位于图像边缘和细节处的噪声像素都可以被区分出来.在图像滤波阶段,算法引入了双阈值迭代方法对确认出的噪声像素赋值,提高了对噪声像素的估值精度,从而有效地保护了图像的细节.无论是主观视觉评估还是客观数据评估,实验结果都表明了该算法优于现有的很多方法.
In order to remove the impulsive noise in the image and effectively protect the edge and detail of the image, a new two-phase random noise impulse noise filter with random details is proposed in the image interfered by random impulse noise.In the noise detection stage, Aiming at the problem that the edges and detail pixels in the image are difficult to distinguish effectively from the noise pixels, an absolute grade difference statistic (S-ROAD) based on S-estimation is proposed. By introducing the S-estimate to the ROAD statistics, By the S-ROAD statistics, most of the noise pixels in the image, including the noise pixels located at the edge and detail of the image, can be distinguished.In the image filtering phase, the algorithm introduces double The iterative thresholding method assigns the identified noise pixels and improves the accuracy of the noise pixel estimation, thus effectively preserves the details of the image.Experimental results show that the proposed algorithm is superior to the existing There are many ways.