异型纤维图像去噪分析

Analysis of Denoising in Shaped Fiber Image

为了去除异型纤维图像中的噪声,首先分析了异型纤维图像中的噪声模型,然后针对噪声模型提出了一种能同时去除异型纤维图像中高斯和脉冲混合噪声的去噪算法.该算法在全变差(Total Variation,TV)算法的基础上进行了算法改进,综合了中值滤波的优点,在达到去噪目的的同时,较好地处理了去除噪声、保留边缘细节信息这对在图像去噪中存在的矛盾.同时,对参数的选取也做了分析,较好地平衡了去噪效果和处理效率问题.数值对比实验中的视觉效果和客观标准均表明了该去噪算法的有效性。

After analyzing the noise model in the shaped fiber image, a denoising scheme is presented to smooth the Gaussian noise and impulsive noise in the image. This scheme is based on Total Variation algorithm and median filtering, can keep the information of fiber edges while denoising. Meanwhile, based on the threshold of the parameters, the efficiency of the algorithm and the effects of denoising is well balanced. The experimental result shows the good performance of the proposed algorithm.