线激光扫描技术多用于零件的表面检测，也可用于物体的三维重建。许多浮雕工艺品由于没有数字模型而无法复现，可利用线激光扫描将浮雕逆向生成三维模型，便于浮雕工艺品的加工。通过结合机器人与线激光，获取浮雕的点云数据，逆向建立浮雕的三维数字模型。搭建三维重建系统，根据实际模型的尺寸计算出机器人的扫描路径，机器人结合线激光扫描得到浮雕点云数据；进行点云数据的预处理，再根据基准平面，补偿带有机器人误差的点云数据；利用基于衍生的迭代最近点（GICP）算法进行点云的自动拼接，并对点云进行后处理；随后利用Delaunay三角剖分算法与曲面重建算法进行三维模型的重建。以老鹰浮雕为重建对象进行实验，实验结果表明用路径扫描拼接的点云处理更为方便，补偿数据后，精度可提高40.48%，有显著补偿效果，浮雕模型重建后与基准模型的标准差平均为0.0576 mm，可满足浮雕模型的重建需求。

Laser line scanning technology is primarily used for surface inspection of mechanical parts and three-dimensional (3D) reconstruction of objects. Many relief artifacts cannot be reconstructed owing to the absence of digital models; however, laser line scanning can perform relief reversal and generate a 3D model for the processing of these relief artifacts. In this study, a robot and line laser are combined to obtain point cloud data and reconstruct a 3D digital model of reliefs. To calculate the scanning path of a robot, a 3D reconstruction system based on the size of the reference model is constructed, and the robot is combined with the laser line scan to obtain the point cloud data of the relief. The point cloud data is preprocessed, and then the point cloud data with robot error is compensated according to the reference plane. Based on the derived iterative closest point (GICP) algorithm, the point cloud is automatically stitched and postprocessed. The 3D model is then reconstructed according to the Delaunay triangulation and surface reconstruction algorithms. The experiments were performed with an eagle relief as the reconstruction object. The experimental results show that implementing path scanning in point cloud stitching provides an easy to execute process, with 40.48% improvement in the accuracy after data rectification and significant error compensation. Additionally, the average standard deviation between the reconstructed relief and reference model is 0.0576 mm, meeting the requirements of the reconstruction effort.