基于未标定图像的手术场景的三维重建  

3D reconstruction of a surgical scene from two uncalibrated images

作　　者：陈静[1] 罗述谦[1] 

机构地区：[1]首都医科大学生物医学工程学院,北京100069

出　　处：《北京生物医学工程》2007年第2期129-131,135,共4页Beijing Biomedical Engineering

基　　金：国家自然科学基金(60472020)资助

摘　　要：利用一个塑料人体模型模拟病人躺在手术台上模拟手术场景和手术对象,并从不同视角拍摄两张照片。使用MAPSAC方法估计基础矩阵,并用自标定方法恢复相机的投影矩阵。实验结果表明此方法可以在近似欧氏空间中从两幅图像重建外科手术场景和手术对象。To used a plastic human body model to simulate a patient lying on operating table. Then two photos of the virtual patient and surrounding objects were taken from two views. MAPSAC method was used to estimate the fundamental matrix, and self calibration was used to recover the camera projective matrix. The results showed that this method could be used to reconstruct the surgical scene in a close Euclidean soace.

关 键 词：增强现实 对极几何 基础矩阵 自标定 

分 类 号：R318.04[医药卫生—生物医学工程]