\r孙 涛¹，闫 巍¹，孙振辉²，张 弢²，宋轶民¹，马信龙²，连宾宾^{1, 3}

AuthorsHTML:\r孙 涛¹，闫 巍¹，孙振辉²，张 弢²，宋轶民¹，马信龙²，连宾宾^{1, 3}

AuthorsListE:\rSun Tao¹，Yan Wei¹，Sun Zhenhui ²，Zhang Tao ²，Song Yimin¹，Ma Xinlong ²，Lian Binbin^{1, 3}

AuthorsHTMLE:\rSun Tao¹，Yan Wei¹，Sun Zhenhui ²，Zhang Tao ²，Song Yimin¹，Ma Xinlong ²，Lian Binbin^{1, 3}

Unit:\r1. 天津大学机械工程学院机构理论与装备设计教育部重点实验室，天津 300354；
2. 天津市天津医院，天津 300211；3. 天津大学仁爱学院机械工程系，天津 301636

Unit_EngLish:\r1. Key Laboratory of Mechanism Theory and Equipment Design of Ministry of Education，School of Mechanical Engineering，Tianjin University，Tianjin 300354，China；
2. Tianjin Hospital，Tianjin 300211，China；
3. Department of Mechanical Engineering，Tianjin University Ren’ai College，Tianjin 301636，China

Abstract_Chinese:\r针对精准外固定支架的设计与开发，提出了交互式外固定支架三维模型的软件系统及外固定支架安装参数的识别算法，解决了传统方法中采用二维影像划线测量获取三维安装参数误差较大的问题．考虑远程化诊疗的潜在需求，软件系统采用浏览器/服务器架构构建基于WebGL三维模型渲染技术的三维交互场景．识别算法分为标记物位置拾取与外固定支架位姿识别两部分，通过区域生长算法分割模型文件，采用最小二乘法识别标记物位置，进而借助标记物与外固定支架的几何关系求解位姿参数．以外固定支架参考环与移动环中心距离为例，开展标记物识别方法与X光影像划线测量方法的实验对比研究．借助高精度测量仪器获取两环中心距离作为标准对照组，将上述两种方法的结果分别与标准对照组进行比较．基于标记物识别方法得到的两环中心距离与标准对照组的结果误差为0.043mm，X光影像划线测量方法的误差为4.903mm．实验结果表明，基于重建软件和标记物识别法获取外固定支架安装参数较传统X光影像划线测量方法精度高，并有效地将精度提升两个数量级，该方法可为开发精准高效可靠外固定支架系统提供技术支持．

Abstract_English:\rAiming at developing an external fixator system with high precision，a software for 3D model rebuilding is designed and the method for obtaining 3D mounting parameters is proposed in this paper that solves the problem of low accuracy in existing systems that apply a 2D image from an X-ray photograph to estimate the mounting parame-ters．Firstly，the browser/server(B/S)architecture is adopted considering future application in long-distance treat-ment．On this basis，a 3D model rendering procedure based on WebGL is established．Then，the mounting pa-rameters can be derived by the 3D model of the external fixator shown in the software．The model is divided into two steps，obtaining the maker position and fixator pose．The marker position is identified by a region growing algorithm on the STL file and with least square fitting．The pose of the fixator is computed by the geometry of the marker and fixator．With the fixator pose available，the mounting parameters，such as the distance between the reference and moving rings，can be easily obtained．Finally，experiments are performed to verify the method．Three groups of experiments are designed to determine the distance between rings：a reference group in which the distance is meas-ured by a high precision measuring device，a 3D model based group and a conventional 2D image based group．In the reference group，200measurements are implemented to measure the center of each ring，and the mean values adopted to construct the coordinates of the centers．The same experiments are repeated 10 times to eliminate the measuring errors．In the 3D model based group，a CT image of the external fixator is converted to STL files and the 3D model is built in our software．The coordinates of the centers of the rings are directly obtained and the distance can be computed．In the 2D image based group，the ellipse and lines are drawn on the 2D image of the external fixa-tor from the X-ray photograph，and the distance is estimated．By comparing the distances obtained by the 3D or 2D model with the measured distance，the error between 3D model based method and measured reference is 0.043 mm，and is 4.903mm for the 2D model based．The results show that our method improves the accuracy compared to the conventional method．This study lays a solid foundation for the design of precise and reliable external fixator systems．

Keyword_Chinese:外固定支架；浏览器/服务器架构；WebGL；区域生长算法；标记物识别

Keywords_English:external fixator；browser/server(B/S) architecture；WebGL；region growing algorithm；marker identification


PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=6467