针对当前逆向工程中对象提取及模型重建效率较低的问题,提出了一种面向室内场景点云的对象重建方法。首先构建直通滤波器,采用改进的RANSAC算法剔除非对象点云,然后根据欧氏聚类提取算法分割出各个对象点云,最后基于α-shape理论批量重建出对象模型。试验结果表明,该方法能够从散乱的室内场景点云中快速、自动地重建出代表真实对象的三维模型,具有较高的实用价值。 Aiming at the problem of low efficiency of object extraction and model reconstruction in current reverse engineering, a method of object reconstruction for indoor scene point clouds is proposed. Firstly, a direct-pass filter is constructed. The non-object point cloud is eliminated by the improved RANSAC algorithm. Then, the object cloud is segmented according to the Euclidean clustering algorithm. Finally, the object model is reconstructed based on α-shape theory. Experimental results show that this method can quickly and automatically reconstruct the three-dimensional model representing the real object from the scattered indoor scene point cloud, which is of high practical value.