摘要
在计算机辅助骨科手术系统中应用增强现实技术能帮助医生准确地定位患者的病灶部位,而视频图像的目标跟踪匹配是实现增强现实的关键技术。针对视频图像匹配中SURF(speed up robust features)特征点性能和匹配效率不足的问题,提出一种改进的基于SURF特征点的FLANN(fast library for approximate nearest neighbors)匹配算法。提取SURF关键特征点,改进其描述符算子,使用改进的FLANN算法进行特征点匹配。通过实验分析比较改进与未改进算法的性能,结果表明该方法的稳定性及快速性较好,具有较强的鲁棒性。
The application of augmented reality technology in computer-assisted orthopedic surgery system can help doctors accurately locate the patient’s surgical site,and tracking targets is the key technology for implementing augmented reality in the frame of video.Aiming at the insufficient problem of performance and matching efficiency of SURF in video image matching,an improved FLANN matching algorithm based on SURF feature points was proposed.The SURF feature point was extracted,whose descriptor was improved.The improved FLANN algorithm was used for feature points matching.The performances of the improved algorithm and the original one were compared through experimental analysis.The improved algorithm has better stabi-lity and rapidity with stronger robustness.
作者
张志敏
李彬
田联房
丁焕文
ZHANG Zhi-min;LI Bin;TIAN Lian-fang;DING Huan-wen(School of Automation Science and Engineering,South China University of Technology,Guangzhou 510641,China;School of Medicine,South China University of Technology,Guangzhou 510006,China)
出处
《计算机工程与设计》
北大核心
2022年第4期941-948,共8页
Computer Engineering and Design
基金
广东省基础与应用基础研究基金项目(2019A1515011148)
中央高校基本科研业务费专项基金项目(2019MS139)
广东省科技计划基金项目(2017A020215164)。
