多视域下基于机器视觉的索力测试

拉索是缆索承重体系桥梁的关键承力或传力构件，在体系构成中的地位至关重要。为解决现有非接触视觉测量存在的测试视场不足、应用场景复杂多变，致使其测试精度与视域范围相互矛盾无法协调统一的技术难题，提出了多视域下基于机器视觉的索力测试新方法。以全息视觉传感器系统获取斜拉索全视域-多视域下的空间几何构型数据，建立了像素映射及图像增强算法计算模型；通过数字空间基准平面的信息转化，对复杂测试场景下视觉传感器光轴射线与被测物体所在承影面非正交情况进行测试参数修正。为避免实测拉索全息性态特征因偏离标准悬链线所导致的索力测量误差，以单一视域下的几何构型拟合降噪后的全息曲线并精确定拉索上下锚固点坐标，进而可由拉索全息性态特征参数精确测算斜拉索索力。试验结果表明：本文方法可在复杂测试场景下量化分析拉索索力计算参数，相较于常规索力测试方法，最大相对误差为9.2%，均方根误差为2.79%，满足工程实践测试精度及稳定性要求，为复杂测试场景下斜拉桥索力监测提供了一种新途径。

Stay Cable Force Measurement Using Machine Vision in Multi-View

Cable is the key component of cable bearing system bridge, which plays an important role in system composition. To solve the existing non-contact optical measurement of the insufficient field and complex application scenarios, resulting in the testing accuracy and the scope of view cannot coordinate the technical problems, a new method for cable force measurement based on machine vision in multi-view is proposed. A holographic vision sensor system was used to obtain the cable spatial geometry configuration data in the full-view and multi-view, and a calculation model of pixel mapping and image enhancement algorithm was established. Through the information transformation of digital space reference plane, the test parameters were corrected for non-orthogonal case of optical axis ray of the visual sensor and the bearing surface of the measured object in the complex test scene. To avoid the cable force measurement error caused by deviation of cable holographic characteristics from standard catenary, the holographic curve after noise reduction was fitted by geometric configuration in a single field of a single view and the coordinates of anchor points up and down of cable were precisely determined, and then the cable force could be accurately measured by the holographic characteristic parameters. The experimental results demonstrated that the proposed method can quantitatively analyze cable force calculation parameters in complex test scenarios. Compared with the conventional cable force measurement method, the maximum relative error is 9.2%, and the root mean square error is 2.79%, which meets the requirements of engineering practice testing accuracy and stability, and provides a new way for cable force measurement of cable-stayed bridge in complex test scenarios.