基于视频亚像素模板匹配算法的索力试验

针对索结构的索力测试，基于室内试验对比了计算机视觉方法和传统加速度传感器方法，探讨了计算机视觉方法的适用性。研究采用伺服静载锚固试验机张拉单根斜拉索钢绞线，简化模拟索张拉受力状态，通过相机摄影的非接触测量方法测量拉索动态响应，结合亚像素模板匹配算法识别拉索动态变形、基频和索力，并与传统的加速度传感器测试结果进行对比。研究结果表明，计算机视觉方法识别的拉索振动频谱峰值明显，基频识别结果与加速度传感器测试结果基本一致，索力识别结果与实测值相比最大误差不超过6%，具备开展实际结构索力测试的能力。

Cable Force Test Based on Video Sub-pixel Template Matching Algorithm

Aiming at the cable force test of cable structure and based on the laboratory test, the computer vision method and the traditional acceleration sensor method are compared, and the applicability of computer vision method is discussed. It is studied the use of the servo static load anchor testing machine to tension a single steel stay-cable strand. The simulation of cable tensioning stress state is simplified. The dynamic response of cable is measured by the non-contact measuring method of camera photography. The dynamic deformation, base frequency and force of cable are identified by the sub-pixel template matching algorithm, and are compared with the test results of the traditional acceleration sensor. The study results show that the rumble spectrum peaks of cable identified by the computer vision method are obvious, and the identifying result of base frequency is basically same as the test result of acceleration sensor. The maximum error between the cable force identification result and the measured value is less than 6%, which has the ability to conduct the actual structural cable force test.