基于边缘特征点匹配的临床X线自动全景成像研究

摘要

目的:本文提出了一种基于边缘特征点匹配的X线自动全景成像方法.方法:一,采用双正交小波变换与Canny算子相结合的算法实现X线图像的有效边缘提取,并通过矩阵计算得到特征点;二,利用Pearson相关系数和随机抽样一致算法来建立图像之间的相关性和匹配性;三,利用渐入渐出的融合算法实现X线无缝全景成像.结果:该算法可快速高质量地实现骨科临床无外部特征的X线图像序列全景拼接.结论:该算法对弱对比度的X线图像序列全景拼接具有很强的鲁棒性,允许X线图像序列之间存在较小的图像旋转和缩放差异,有助于减少X线使用剂量,具有较强的应用价值.%An automatic clinical X-ray panoramic imaging method based on the matching of feature points on the edge is presented in this paper. Methods: First, the effective edge of the X-ray image can be extracted by Canny operator combined with biorthogonal wavelet transform. And the feature points of the image can be obtained by matrix calculation. Second, the correlation and the match degree of the images can be established found by using Pearson product-moment correlation coefficient algorithm and the random sample consensus algorithm. At last, the X-ray seamless panoramic imaging can be achieved by using fade-in and fade-out fusion algorithm. Results: The high-quality clinical X-ray panoramic imaging of orthopedics can be obtained fast by using this method without external characteristics. Conclusions: This algorithm is robust for the weak-contrast X-ray image sequence. It allows small image rotation and scaling difference between the X-ray image sequences and it reduces the x-ray dose. Therefore it may be applied widely.