Improved Reduced Convolution for Minkowski Operations in Laser Nesting Application
Qirui Hu,Zheping Zhou,Zhiwei Lin,Jianzhong Fu,Congcong Luan
DOI: https://doi.org/10.2139/ssrn.4234099
2022-01-01
Abstract:Nesting in laser cutting application is required to place as many parts as possible on a given metal sheet. One key of the nesting algorithm is to detect the overlaps between parts. The No-Fit Polygon, also known as the Minkowski difference method, is considered to be the general solution of the overlap detection problem. However, calculating the Minkowski difference of the two arbitrary polygons is never an easy task. So far, several approaches have been proposed, but only two are practical: the brutal-force and the convolution. The latter is considered to be more promising with the concept of reduced convolution, in which, however, the accurate boundary extraction with degenerate edges or points still remains a problem. In this paper, a new accurate boundary extraction algorithm is proposed for laser cutting nesting. The potential segments of Minkowski difference is calculated with the reduced convolution method. Then boundary quadrangles and triangles are constructed based on the potential segments. Together with input polygons, the quadrangles and the triangles are joined with the Boolean union operation. In this way, the error-prone boundary extraction problem is well solved. Furthermore, through polygon offset and Boolean operation, proposed method can also find the degenerate edges and points. The proposed algorithm takes the advantage of the reduced convolution method; therefore, it is fast; meanwhile, it is more stable. The efficiency and robustness of proposed algorithm is verified by several comparative examples. Also, the feasibility is verified by actual laser nesting experiments with exist method.