基于图像处理技术的X线图像拼接方法研究  

The Research of X-ray Image Stitching Method Based on Image Processing Technique

作　　者：杨帆[1] 何艳[1] 邓振生[2] 严昂[3] 

机构地区：[1]贵州医科大学生物医学工程教研室,贵州贵阳550004 [2]中南大学生物医学工程研究所,湖南长沙410083 [3]中南大学湘雅二医院设备科,湖南长沙410011

出　　处：《贵阳医学院学报》2016年第4期395-398,401,共5页Journal of Guiyang Medical College

基　　金：国家自然科学基金地区科学基金项目(81460206)

摘　　要：目的:探讨一种基于图像处理技术的全脊柱及下肢X线图像自动拼接算法的可行性。方法:采用医学图像处理技术,通过相位相关算法确定粗略重叠区域,利用Harris算法提取特征点和角点邻域灰度信息实现图像的精确配准,对40组多次拍摄的脊柱或下肢X线图像进行算法拼接,并与设备自带软件拼接的图像进行比较。结果:对40组多次拍摄的脊柱或下肢X线图像进行算法拼接,误差4 mm内拼接成功率达到95%,高于自带软件拼接结果。结论:算法拼接法能有效实现多次拍摄的脊柱或下肢X线图像拼接。Objective： To explore the feasibility of automatic stitching method for full-spine and lower limbs X-ray images based on image processing technique. Methods： Medical image processing technology was adopted in this research. Phase correlation algorithm was used to identify rough areas,and Harris algorithm to extract features and corner neighborhood gray scale information to actualize accurate image registration. 40 pairs of spine or lower limbs X-ray images were used to conduct algorithm stitching and compared with the result of stitching software. Results： 40 pairs of spine or lower limbs X-ray images were used to conduct algorithm stitching,and the stitching success rate with deviation within4 mm was 95%,which was higher than that of stitching software. Conclusion： The algorithm stitching method can effectively actualize image stitching of spine or lower limb X-ray images.

关 键 词：X线图像 相位相关 角点 自动拼接 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]