This paper reports an algorithm to chop point cloud into layer-wise slices for additive manufacturing. It starts with intersecting slicing plane with the 3D input points, generating planar samples. Then, an online learning model, known as competitive segments representation (CSR), extracts their implicit topology and distribution. CSR structure is a restricted graph that equals to multiple polylines, which are meanwhile piecewise linear approximation to the principal curves of samples. Edge segments of CSR compete with each other for representing consecutively given samples. They dynamically move, grow, shrink or rewire subject to several heuristic rules. Those rules are designed to depress abnormal data, enable lifelong learning, recover salient feature and ensure correct topology. Assembling them together allows online tracking of changing curves. Once CSR converges on one slice, learnt curves are reused as initial estimation for the next. By this practice, shape coherence of successive slices is efficiently utilized, and the ongoing learning output all subsequent slices incrementally. We have verified the feasibility of proposed algorithm both on synthesized data and scanned points.
Additive Manufacturing Curve reconstruction Topology learning Principal curve
This is a preview of subscription content, log in to check access.
This work was supported by the National High Technology Research and Development Program of China (Grant No. 2015AA042502). We are grateful to Xu Jinting for providing the PLSP implementation.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
Zhang L, Dong H, Saddik AE (2016) From 3d sensing to printing: a survey. ACM Trans Multimed Comput Commun Appl 12(2):27Google Scholar
Gao W, Zhang Y, Ramanujan D et al (2015) The status, challenges, and future of additive manufacturing in engineering. Comput Aided Des 69:65–89CrossRefGoogle Scholar
Mohan Pandey P, Venkata Reddy N, Dhande SG (2003) Slicing procedures in layered manufacturing: a review. Rapid Prototyp J 9(5):274–288CrossRefGoogle Scholar
Fritzke B (1994) A growing neural gas network learns topologies. In: Proceedings of the 7th international conference on neural information processing systems. MIT Press, Cambridge, MA, USA, pp 625–632Google Scholar