A novel versatile method for generating machining path directly from point cloud
Wenxing Guo,Xiaokang Huang,Bowen Qi,Xukai Ren,Huabin Chen,Xiaoqi Chen
DOI: https://doi.org/10.1016/j.jmapro.2024.04.019
IF: 5.684
2024-04-17
Journal of Manufacturing Processes
Abstract:3D vision technology, as a primary means of acquiring point clouds, emerges as a promising solution for manufacturing, captivating both industry and academia. Point clouds offer an intuitive depiction of the workpiece surface and are extensively explored for the generation of machining paths. However, existing researches on point cloud path generation usually focus on specific processing method, often neglecting the development of versatile methods. In response, this paper proposes a versatile approach based on Weighted Locally Optimal Projection (WLOP) for the direct generation of machining paths from point clouds, which can be easily utilized in diverse machining scenarios. The method begins by introducing the WLOP operator, strategically reducing scattered points sampled from the Region of Interest (ROI) to a 1D geometry, thereby facilitating path generation. Then, the challenge of path point serialization is reformulated into a minimum spanning tree problem and is subsequently addressed using the Prim algorithm. Further refinement is achieved through the application of a pruning algorithm designed to eliminate small branches. To enhance the practicality of the generated paths, a path smoothing objective function is formulated. Curvature constraints are then introduced to promote smoother paths. Experimental results validate the method's high accuracy and high versatility, showcasing its superiority to various existing methods. The method provides a simple and fast way to generate accurate tool paths based on point cloud data in diverse machining scenarios.
engineering, manufacturing