【摘要】 针对现有手术导航系统中存在不同三维空间注册精度低、需要额外复杂设备辅助完成手术器械实时跟踪的问题,设计并实现一种基于视觉的手术导航系统。利用正弦结构光重建病人空间面部点云;利用人脸面部特征进行病人空间与模型空间由粗到精的空间注册;利用分步角点检测算法和双目立体视觉原理恢复手术器械上标记点的三维坐标,并利用卡尔曼滤波获取手术器械精确的位置与姿态。实验表明,基于视觉的手术导航系统能够实现准确、稳定、实时的手术器械跟踪要求。更多还原

【Abstract】 A navigation system based on vision was designed to solve the problem of using complex equipment to assist the real-time tracking of surgical instruments because of low registration of 3D in the existing surgical navigation system. Face point cloud was reconstructed in the patient’s space by using the sinusoidal structured light; From rough to precise registration, patient space and model space were made by using facial feature; The three dimensional coordinates of the marker points on surgical instruments was restored by using continuous corner detection algorithm, binocular stereo vision principle and Kalman filter algorithm to obtain accurate position of surgical instruments. The experiments show that the surgical navigation system based on vision can achieve the requirements of accurate, stable, real-time tracking of surgical instruments.更多还原