A new algorithm of gauge calculation based on laser camera detection was proposed. This algorithm, according to track gauge defined, finds the minimum distance in 16 mm between two rails work side under rail top tread as track gauge measuring value. Considering the influence of camera lens distortion, a nonlinear calibration model of spatial attitude relationship between two laser camera sensors was established, and a calibration program was proposed. In order to quickly locate the profile feature points, this paper put forward the concept of the rail profile feature curve, and proposed a definition of the characteristic curve. Using feature curve can rapidly find the profile feature points. Compared with the traditional detection method, the new algorithm only uses two laser camera sensors, and does not need other auxiliary sensors to modify the car body vibration. Finally, by setting obstacle test through the actual subway line, it is approved that this detection method has the characteristics of certain anti-interference ability, high stability, simple structure, strong real-time online calculation, high detection accuracy and so on.%提出一种基于激光摄像检测的轨距计算新算法,该算法将按照轨距定义寻找轨顶踏面下16mm内2股钢轨工作边之间的最小距离作为轨距测量值.考虑摄像机镜头畸变影响,建立2个激光摄像式传感器之间空间姿态关系的非线性标定模型,并给出一种标定方案.为了能快速定位轮廓特征点,提出钢轨轮廓特征曲线的概念,并给出特征曲线的一种定义方式,利用特征曲线能快速寻找轮廓特征点.与传统检测方法相比,新算法只用了2个激光摄像式传感器,不需要其他辅助传感器来对车体振动进行修正.最后通过实际地铁线路设置障碍试验,验证了该检测方法具有一定的抗干扰能力、稳定性高、结构简单、在线计算实时性强和检测精度高等特点.
展开▼
