摘要
根据校准靶标志点影像的特点,提出一种自动的离线C形臂相机模型校准方法。应用聚类算法实现标志点信息的自动识别,通过行列索引表简化标志点网格姿态的识别,自动建立标志点在影像坐标系和空间坐标系下坐标的对应关系,经几何变形校正后,实现C形臂相机的校准。实验数据证明,该方法具有良好的鲁棒性,误差小于1mm,可应用于基于术中C形臂X线影像的手术导航系统。
According to the analysis of characteristics of calibration target marker images, an automated off-line calibration method for C-arm camera model is proposed. The method can extract the markers information from images automatically applying clustering algorithms. Through the row and column index table, the problem of the orientation recognition of the marker mesh is simplified and the correspondence between markers coordinates in image coordinate system and those in spatial coordinate system can be established automatically. Basing on this correspondence, the geometric distortion can be corrected and the C-arm camera model calibration can be performed. This automation method is proved to be robust by experimental results and the error is less than 1 mm. The method can be applied in the surgical navigation system based on operative C-arm fluoroscopy.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第15期7-9,12,共4页
Computer Engineering
基金
上海市科委基金资助重点项目"多系列手术专用软件系统研发及其骨科应用"(045115002)
关键词
C形臂
校准靶
自动识别
手术导航
C-arm
calibration target
automatic recognition
surgical navigation
