摘要
提出了一种基于机器视觉的轨距点定位算法能较好地适应并解决这一问题,该算法在基于机器视觉的基础上获得轨道断面光带图像,并结合LabVIEW软件对图像进行处理得到轨道截面轮廓的测点数据;其次,对轨头测点数据进行曲线拟合得到轨道顶面曲线,应用冒泡算法排序得到轨头测点到顶面曲线切线平移值的距离,选取距离较小的值做二次曲线拟合,求取曲线与切线平移后直线的交点达到对轨距点定位;最终实现对轨距的准确检测。试验结果表明:该算法能够实现轨道轨距点的快速精确定位,捕捉精度可达到1 mm,有效地对轨道轨距进行了检测。
The gauge irregularity is the bottleneck for the security of high-speed train. The problem is resulted directly by the problem that gauge factors point can not be precise positioned. This paper presents a point-based machine vision gauge positioning algorithm. It can solve this problem effectively. The algorithm is based on machine vision. When the image is obtained by the light rail sections, LabVIEW is used to process the image and obtain the measuring point data of track sectional profile. The top surface of the track curves is obtained by fitting the curve of rail head measuring point data ; The bubble sort algorithm is applied to measure tangent curve shift values of top surface of the rail head ; A smaller distance value is determined to complete quadratic curve fitting, such that the tangent line can reach translational gauge point positioning; Ultimately, accurate detection of the gauge can be achieved. The test results show that： the algorithm can achieve fast and accurate positioning track gauge points, and the precision can reach 0.07mm. It effectively tracks gauge tested.
出处
《实验室研究与探索》
CAS
北大核心
2015年第2期122-124,131,共4页
Research and Exploration In Laboratory
基金
甘肃省自然科学基金项目(121RJZA046)
