High accurancy calibration is the precondition of the measurement by structured light.To achieve the above purpose,this paper presents a calibration algorithm for line structured light based on stereo vision.Firstly,intrinsic parameters and the transformation between stereo cameras were calibrated by the Tsai method.Then,laser pixels extracted from stereo image pairs were matched by epiploar constraint.At last,the matched points were reconstructed to 3 dimensions to achieve the calibration of laser plane.Compared with other popular methods,the proposed laser plane calibration algorithm needs no calibration target.In contrast,dense calibration points can be got by projecting the structured light to any object.The algorithm′s availability is assessed by its application to a self-referenced line structured light scanning system with two cameras.Experimental results show that the RMS calibration error by the proposed algorithm is less than 0.034 mm and the scanning error is 0.065 mm.