Rail profile detection is a key and difficult issue for rail geometry parameters detection. Laser camera sensor is calibrated by 2D planar drone to build calibration model of world coordinate system and image coordinate system. Experimental calibration is carried out by checkerboard

300 groups of datas are obtained

and non-linear least square method is used to calibrate model parameters. The results show that the measurement error of the horizontal and vertical of checkerboard can be controlled around 0. 3 mm by non-linear least-square method

it can be well applied for rail profile measurement.