基于SIFT的软骨切片电镜图像拼接算法  

Electron Microscope Cartilage Slicing Image Stitching Algorithm Based on SIFT

作　　者：刘纪红[1,2] 张倩[2] 郭晴[2] 

机构地区：[1]东北大学中荷生物医学与信息工程学院,辽宁沈阳110819 [2]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《东北大学学报（自然科学版）》2014年第6期785-789,共5页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(61173028)

摘　　要：软骨组织断层的电镜图像拼接是人工软骨组织培养过程中的关键技术之一.对二维断层电镜图像的拼接方法进行了深入研究,利用SIFT算法对图片进行关键点检测,利用kd-tree建立数据索引,进行关键点的快速匹配,确定图像位移,实现了局部断层图像的无缝拼接,有效解决了电镜图像视角小的问题.在VS2010环境下利用C语言进行仿真,实验表明,该方法具有鲁棒性强、实时性好的特点.证明该方法可以为人工软骨组织的培养提供客观的科学支持.During the establishment of simulation platform, the electron microscope cartilage tomographic image stitching is one of the key technologies. The two-dimensional electron microscope method was studied. Firstly, the SIFT algorithm was used to detect key points in the image, and the kd-tree was used to create data index to match the key points fast, then the image shift could be determined. Finally, the seamless stitching of partial slicing image was realized and the problem of small angle for electron microscope image was solved effectively. C language was used in VS2010 environment for simulation. The experimental results showed that the method presented is robust and real-time. It was proved that this method could provide objective scientific support in the cultivation process of artificial cartilage tissue.

关 键 词：软骨 电镜图像 图像拼接 SIFT kd—tree 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]