基于OpenCV的针穿刺靶点动态跟踪算法及其实现  

A novel algorithm base on OpenCV for tracking dynamic targets in needle insertion

作　　者：徐飞 高德东 王珊 阿占文 

机构地区：[1]青海大学机械工程学院,青海西宁810016

出　　处：《青海大学学报（自然科学版）》2017年第6期49-55,88,共8页Journal of Qinghai University(Natural Science)

基　　金：国家自然科学基金项目(51665049);青海省自然科学基金项目(2015-ZJ-906);青海大学中青年科研基金项目(2016-QGY-3)

摘　　要：为了降低针穿刺软组织过程中内部靶点运动对手术精度和虚拟训练系统中路径规划的影响。文中在提取假体内部动态靶点特征的基础上,选取包含特征标识物靶点的矩形作为感兴趣区域,设置网格线条对靶点中心进行跟踪,提出了一种基于OpenCV的组织内部靶点特征提取和动态跟踪算法,在Visual C++环境下利用OpenCV计算机视觉库实现了算法功能。通过在PVA假体组织内嵌入标识物靶点对算法进行验证,实验结果表明算法动态跟踪靶点成功率达到95%以上,具有较高的跟踪效果和实用性,可实现对针穿刺过程中靶点的动态跟踪。In order to reduce the effect of target motion on the accuracy of surgery and the path plan-ning of virtual training system in the process of needle insertion, the features of dynamic targets are extracted with the image morphology, the rectangles containing the feature of moving targets are se-lected as a region of interest, and the grid lines are defined to tracing the center of target, and a no-vel algorithm based on OpenCV is provided to trace the dynamic targets inside the soft tissue. The proposed tracking algorithm is designed and validated by embedding marker targets into the phan-tom, which is used as a free computer vision library OpenCV under the compiling environment of Visual C＋＋ . The marker targets are embedded into the PVA phantom for verifying the provided al-gorithm , and the experimental results show that tracking success rate of algorithm is over 9 5 % , with high tracking effect and practicability. The dynamic tracking of the target can be achieved.

关 键 词：针穿刺 软组织变形 图像处理 OPENCV 

分 类 号：TN911.73[电子电信—通信与信息系统]