Abstract:
Considering the problems of point cloud defects of reflective workpiece through three-dimensional scanning process and long processing time of the point cloud, this paper proposes a point cloud consolida- tion algorithm to improve the quality of the scanned three-dimensional data and to reduce its processing time. The algorithm can thus improve the accuracy and efficiency of the three-dimensional reconstruction. The first step of the algorithm was to acquire the unorganized point cloud with nonuniform distribution, outliers and noise, then a kD-tree based resampling method was used to subdivide the points into multiple subsets. For each of the point subset, noise reduction was performed through an improved local optimal projection algorithm. Finally, the edge-aware resampling method was adopted to upsample the points to preserve and consolidate the feature of the point cloud. Experimental results show that the point cloud consolidation algo- rithm is effective.