粗糙集在图像边缘增强滤波中的应用

为使图像噪声滤除的同时,边缘细节信息能得到有效地保护,基于粗糙集理论介绍了一种新的边缘增强滤波方法。首先,基于粗糙集不可分辨等价关系划分,分离出噪声点和非噪声点,对噪声点通过中值滤波进行滤除,然后,通过粗近似精度和方向模板检测边缘的连续性和方向,以具有最大粗近似精度的模板的灰度均值取代中心像素点灰度。在所有进行边缘检测的滤波算法中,该算法是唯一边缘测度在多次迭代运行后不会减小的滤波方法,同时通过对不同噪声程度的椒盐噪声和高斯噪声的滤波实验,说明该方法在有效滤除噪声同时能使边缘细节得到保护和增强,且比其它传统的空域和频域滤波方法具有更好的噪声适应性。

The Application of Rough Set to Edge Enhancing Filtering

An edge enhancing filtering method based on rough set is introduced here to remove noise,while the detail edge information of the image can be preserved.First,based on the partitions of indiscernibility equivalent relation,noise pixels can be separated and be removed by median filtering.Next,by using rough approximation precision and direction templates,the continuous and edge direction are detected.Then,the intensity of the central pixel is substituted for the average intensity of the direction template that has maximum rough approximation precision.In all of the filtering algorithms used to detect edge,this is the only algorithm that the edge measure will not decrease after implementing iteratively.In the experiment of filtering salt & pepper noise and Gaussian noise,the method can remove the noise efficiently as well as the detail edges are preserved and enhanced.It also has better adaptability than conventional space domain or frequency domain methods for filtering.