Storage and Reconstruction Method of Sparse Matrix for 3D Printing Based on Irregular Block Compression

The current existing 3D Printing(3DP) is commonly characterized as servo motion point by point which reduces forming efficiency. So, it develops towards high efficiency and high precision, such as Digital Light Processing(DLP), Selective Laser Melting(SLM), section-by-section printing, etc. Generally, in order to improve the printing accuracy, it is necessary to rasterize the cross section connected domain with higher resolution whose subsequent optical conversion and other steps will generate more metadata at the same time. This in turn exponentially increases the data volume of slicing devectorization lattice to large scale, and limit the part size directly. Therefore, a storage and reconstruction method of sparse matrix for 3D printing based on Irregular Block Compression(IBC) is proposed. Firstly, in the original Model Coordinate System(MCS), the layered multi-connected domain is obtained via 3D Axis-Aligned Bounding Boxes(AABB) and the Layered Cross-sectional Mask(LCM) is generated. The rasterized lattice is constructed according to the preset resolution and converted to sparse matrix. The probabilistic mathematical expectation of rectangular Regular Blocks(RB) as independent events is calculated according to sparsity. The calculation proves that Irregular Blocks (IB) occupy major share in the sparse matrix. Therefore, the storage and reconstruction method should be optimized in primary consideration of the irregular blocks. Innovatively, the concept of irregular block of sparse matrix is proposed. Aiming at irregular connected sparse characteristics of sparse matrix, the connected nonzero block between adjacent rows are combinatorially storaged to construct inter-connected irregular block. The value of non-zero elements and their valid location information is saved by lossless compression to obtain the first row index, the first column index, consecutive amount and value set. The data recovery and reconstruction of 3D printing layered section is carried out based on irregular blocks. The solid model is printed with multi-layer continuous surface by calculating the similarity of two adjacent sections. Taking in-line engine block and multi-genus porous revolution ring as examples, compared with traditional Compressed Row Storage algorithm(CRS) and Block Compressed Row Storage algorithm(BCRS), in terms of storage amount improvement, IBC is better than CRS with 80.60%, IBC is better than BCRS with 14.62%, which effectively reduces the time complexity of the algorithm. In terms of storage space, IBC is better than BCRS with 22.56%, which effectively reduces the space complexity of the algorithm. The IBC is especially useful for 3D printing of complex shape models whose layered section is region connected large-scale sparse matrix. © 2020, Science Press. All right reserved.