Rail wear dynamic measurement techniques are important to ensure the security of railway system and the efficiency of maintaining. Nowadays, a widely used approach is based on structured-light vision sensor, which includes a camera and a laser projector generating a light plane on the inner side of the rail. Considering the sensor installation error and vibration of running train, it is difficult to ensure the perpendicularity between the light plane and the rail, which leads to large measurement error. In order to overcome the problem, a motion deviation rectifying method is proposed. Rail section profiles are obtained by a multi-line structured-light vision sensor with multiple parallel light planes. Each profile can be segmented into two curves. One representing the rail waist is used to fit the rail longitudinal axis (parallel to train running direction) and establish an auxiliary plane perpendicular to the direction. Deviation rectifying process is to project the rail section profiles onto the auxiliary plane to recover perpendicularity between the light plane and the rail. Then another curve representing the rail head is used to calculate the rail wear based on the standard profile. The experimental results indicate that the method reduces the rail wear measurement error, meeting the requirements for real-time dynamic measurement of rail wear.
Read full abstract