摘要
双目立体视觉技术能够测量物体的三维信息,将图片信息转换为物体的三维坐标值。双目视觉在穴位的坐标测量中具有非接触、实时性、精确度高和较好的鲁棒性等特点。基于双目立体视觉原理,通过MATLAB标定工具箱对两相机的内外参数进行标定,并通过OpenCV对图像进行极线校正和立体匹配,从而得到视差图和穴位的三维坐标值,并根据手臂和针灸机器人坐标变换关系与齐次变换矩阵将手臂坐标系下的坐标值转换为针灸机器人坐标系下的坐标,为机器人的针灸操作进行定位。通过初步的机器人定位针灸实验表明,该方法的测量精度在允许的范围之内。
Binocular stereo vision can perceive the three-dimensional information of the object and transform image information into three-dimensional coordinates of objects. transform image information into three-dimensional coordinates of objects. Binocular vision has the characteristics of non-contact, real-time, high accuracy and good robustness in the coordinate measurement of acupoints. Based on the principle of binocular stereo vision, this paper calibrates the internal and external parameters of the two cameras through the MATLAB calibration toolbox, and performs polar line correction and stereo matching through OpenCV to get the parallax map and the three-dimensional coordinate values of the acupoints. The coordinate values in the arm coordinate system are converted into coordinates in the coordinate system of the acupuncture robot, and the robot′s acupuncture operation is positioned. The robot positioning acupuncture experiment shows that the accuracy of the method is within the allowable range.
作者
王雪军
徐天成
张小强
宋晨阳
林祺
Wang Xuejun;Xu Tiancheng;Zhang Xiaoqiang;Song Chenyang;Lin Qi(National Key Laboratory of Transient Physics,Nanjing University of Science and TechnoIogy,Nanjing 210094,China;Nanjing University of Chinese Medicine,Nanjing 210023,China)
出处
《电子测量技术》
2018年第22期66-70,共5页
Electronic Measurement Technology
基金
国家级大学生创新训练计划(201610315011)
江苏省研究生实践创新计划(SJCX17_0406)项目资助
关键词
双目视觉
相机标定
立体匹配
穴位测量
binocular vision
camera calibration
stereo matching
acupoint measurement
