In order to improve the calibration precision of the cross-structured light sensor, a cross structured-light calibrating method is proposed. A two-step algorithm for extracting the straight line feature of the imaging light strip and fit the pixel straight line equation in the pixel coordinate system. Through coordinate conversion, get The image line equation in the camera coordinate system is combined with the camera origin to find the plane equation. This plane equation is combined with the two-dimensional target plane equation in the camera coordinate system to solve the laser line equation in the camera coordinate system. Change the target posture, obtain multiple sets of intersecting light line information, and fit the cross laser plane using the principle of random sampling consistency. Finally, the three-dimensional reconstruction of the ring gauge standard parts and error analysis are carried out. The experiment measured the average value of the inner radius error of the ring gauge is 0.155 8 mm. This method makes full use of the point information on the straight line feature, solves the problem of fewer feature points in the previous fitting light plane algorithm, enhances the robustness of the algorithm, and has a certain practicability.