为去除计算层析(CT)图像中的环形伪影，提高CT图像的定性分析能力和定量测量精度，提出了一种改进的Canny算法用于环形伪影校正。首先采用S-L滤波器对原始投影数据进行滤波，增强伪影信息。接着，设定角度阈值对梯度方向进行限制，排除斜角方向边缘的检测，实现竖直方向边缘点检测。然后，设置梯度阈值和链长度阈值实现伪影边缘点的检测与连接。最后，采用分段B样条拟合法对投影数据进行校正，实现多个连续的环形伪影去除。对含有环带伪影的实际CT图像进行了校正实验。结果表明，校正后CT图像环形伪影去除干净，且细节区域的标准差基本没有变化，均匀区域标准差减小，信噪比增益达2.182 dB。该方法既可以校正多个相连的环形伪影，也可以校正单个分散的环形伪影，同时还可较好地保持图像细节和分辨率。

An improved Canny algorithm is proposed to remove ring artifacts in Computerized Tomography (CT) images and to improve their qualitative analysis ability and quantitative measurement precision. S-L filter is adopted to filter original projection data and enhance artifact information before using Gaussian filter. Then, the angle thresholds are set to limit gradient direction, and to exclude diagonal direction edge detection and implement vertical direction edge detection. Furthermore, a gradient threshold and an angle threshold are set to detect and connect the marginal point of artifacts. Finally, the piecewise B-spline fitting method is used to correct the projection data and eliminate multiple continuous ring artifacts. Correction experiments are carried out for the actual CT images including ring artifacts. The results demonstrate that the proposed approach can remove all of the ring artifacts of reconstructed images, keep the standard deviation in detail areas basically unchanged, reduce the standard deviation in uniform areas and increase the Signal to Noise Ratio(SNR) by 2.182 dB. The proposed method is effective in correcting not only the multiple continuous ring artifacts but also the single dispersed ring artifacts, at the same time the CT image details and resolution are reserved.