When a line laser scans metal workpieces， the high reflectivity of certain metal surfaces often leads to poor scanning results. While grid laser scanning is faster than single-line and multi-line scanning， challenges such as center extraction and optical plane positioning hinder its ability to meet the demands of efficient real-time detection of metal workpieces. This study proposes a novel method to address these issues by calibrating the rotation angle of a polarizer to optimize the suppression of high surface reflectivity through the polarization state of the grid laser. In addition， a laser center mapping technique based on normal vectors is developed， where the laser center is mapped to the optical plane using the eigenvector corresponding to the maximum eigenvalue of the pixel Hessian matrix. A center extraction method based on affine transformation is also introduced， enabling the transformation of non-orthogonal grids into orthogonal ones to improve the accuracy of 3D reconstruction. Furthermore， an error correction model is established to reduce measurement system errors by accounting for variations in height. Experimental results demonstrate that the traditional grid laser algorithm struggles with missing edges when scanning highly reflective metal surfaces， resulting in a measurement error of 0.8 mm. In contrast， the proposed method effectively suppresses the impact of high reflectivity， achieves accurate mapping between the laser center and the optical plane， and reduces the measurement error to 0.39 mm. After error correction， the accuracy is further improved， achieving a measurement error of just 0.1 mm， representing an 87.5% reduction compared to the traditional grid laser algorithm.In conclusion， this study investigates 3D reconstruction using grid laser scanning under high-reflection suppression for metal surfaces， significantly improving both scanning efficiency and measurement accuracy. The proposed method provides a valuable approach and reference for future industrial measurement applications.