论文部分内容阅读
提出了一种各向异性分段高斯滤波,这种方法除了考虑图像的尺度和方向特性外,还根据边缘区域可分特性,使用了分段滤波的方法来进一步解决边缘保持的问题.该方法通过本文所提出的一种图像噪音方差估计模型来确定图像的基准尺度,并由Hermite变换得到其方向角,然后再通过确定高斯模型的轴向比和自适应尺度,使对选择区域的滤波转变为对分段高斯模型的滤波,从而使计算的可靠性得到增强.从仿真结果可以看出,各向异性分段高斯滤波器在噪音去除和边缘保持的综合性能上要优于其他常用的滤波算法.
An anisotropic piecewise Gaussian filter is proposed in this paper. In addition to considering the scale and direction characteristics of the image, this method uses the segmentation filtering method to further solve the problem of edge preserving according to the separability of the edge region. Through a noise variance estimation model proposed in this paper, the reference scale of the image is determined and its orientation angle is obtained by Hermite transform. Then, the filter region of the selected region is transformed by determining the axial ratio and the adaptive scale of the Gaussian model In order to filter the piecewise Gaussian model, the reliability of the calculation is enhanced.From the simulation results, it can be seen that the anisotropic piecewise Gaussian filter is better than other commonly used filters in noise removal and edge preservation algorithm.