基于形态学滤波的图像伪边缘去除方法

A Method of Removing Fake Edge Based on Morphological Filtering

针对传统边缘检测算子去除由噪声引起的伪边缘难的问题,提出了一种基于数学形态学滤波的图像伪边缘去除方法.该方法首先利用自适应中值滤波器对图像进行滤波,然后用Canny算子对图像进行边缘检测,对处理后的图像进行改进的τ-运算以去除伪边缘,从而得到较理想的图像边缘.仿真实验表明,该方法不仅有效地去除了由噪声引起的图像伪边缘,而且保留了完整的图像边缘,具有较小的均方根误差.

To solve the difficulty in removing fake edge resulting from noise by traditional edge detection operator,a new type of fake edge removing method is proposed on the basis of improved morphological filter. The input image is filtered by adaptive median filter,and the image edge is detected by Canny operator by binary morphological filter. Besides, improved τ-operation is used to remove fake edge. Experiments show that this method can not only efficiently remove pseudo edge as a result of various typed noise, but also preserve integrated edge with less RMSE.