H值特性在钢轨检测中的应用

Application of H-value Characteristics to Rail Detection

钢轨区域提取是基于机器视觉的轨道缺陷实时监测系统的首要关键技术之一,它对铁路自动化检测系统的检测速度、检测精度影响巨大.针对铁路轨道区域结构特点,提出了一种基于机器视觉HSL色彩空间的钢轨区域快速提取算法,将图像转化为HSL色彩空间模型,提取等分线处H值曲线,利用钢轨与道砟交界处H值的突变特性确定轨道边缘点,采用直线拟合确定钢轨边界,实现钢轨区域提取.

Rail region extracting is one of the first key technologies of real-time monitoring system of rail defect based on machine vision. It has a huge impact on speed and precision detection of high speed railway automatic detection system. According to the regional structure characteristics of the railway track,a rapid rail region extracting algorithm is proposed based on machine vision HSL color space. By transforming the collected image into an HSL color space model and extrac- ting H value c,rve at the bisectrix,the track edge point is found according to the mutant charac- teristics of H value at the junction of track and subgrade. Then the rail region extracting is a- chieved by the rail boundary determined by linear fitting.