基于结构光视觉的钢轨磨耗测量方法

分析了基于结构光视觉的钢轨磨耗测量原理,提出一种钢轨磨耗车载动态测量方法.结构光视觉传感器安装在列车底部,测量钢轨内侧横断面轮廓.以钢轨轨腰轮廓作为测量基准,利用最近点迭代(ICP,Iterative Closest Point)算法确定光平面测量坐标系到设计坐标系的旋转矩阵和平移向量,将测量轮廓与设计轮廓对齐,在此基础上计算磨耗值.与已有的方法相比,该方法无需单独设置用于基准测量的视觉传感器,采用同一传感器实现了基准测量和磨耗测量,有效降低了系统成本,操作性强,且无需进行多传感器的全局校准,保证了测量精度.实验结果表明:该钢轨磨耗测量方法具有较好的重复性精度.

Rail wear measurement method based on structured-light vision

The principle of rail wear measurement based on structured-light vision was analyzed. A method for dynamically measuring rail wears in vehicle-mounted was proposed. The structured-light vision sensor was installed at the bottom of the train, and the section profile of the rail was measured. Taking rail waist as measurement benchmark, the rotation matrix and translation vector between light-plane coordinate frame and designed coordinate frame were estimated by iterative closest point (ICP) algorithm, then, the rail waist profile was registered to designed profile, based on which the rail wears were calculated. Compared with previous methods, the proposed method does not need a specially vision sensor to measure the benchmark. Benchmark measurement and wear measurement are achieved with one same vision sensor. System cost is effectively reduced, and it is operable. Measurement accuracy is also guaranteed due to no need of global calibration of multi-sensor. The experimental results show that the repeatability precision of the method is high.