一种基于阈值分割的 CT 图像环状伪影消除方法

摘要

空域内的极坐标变换法及其衍生算法是一种常用的环形伪影消除方法。虽然极坐标法效果好，处理速度快，但是缺点同样明显：处理后的图像会出现明显的模糊，尤其是图像的边缘地带。论文提出的自适应分割算法在极坐标方法的基础上进行了改进，经实验证明，该算法在保证伪影消除效果的同时，最大限度地避免了图像边缘的模糊。同时，算法还会自动跳过无需处理的区域来减小计算量，从而加快图像处理速度。%Polar coordinate transformation method and its derivative algorithm are a common method of ring arti -fact reduction .Although the effect of this method is good and fast , but the disadvantage is equally obvious:the processed image will appear obvious blurring , especially in the fringe of the image .The paper presents an adap-tive segmentation method which based on the polar coordinates transformation method .The experiment proved that this method can provide a better processed image , which has a clearer fringe .The method will also auto-matically skip the area without valid data to reduce the amount of computer 's calculation , so as to speed up the image processing speed .