A team of researchers has developed two new algorithms for processing point clouds that can significantly improve the quality of 3D scans. The method is particularly aimed at the precise capture of edges and complex structures.
When objects are digitized in 3D, the scan goes through several processing steps. After the actual recording by scanners – whether by photogrammetry, laser or structured light – point clouds are created first. These three-dimensional data sets must then be analyzed in order to identify surfaces, curves and edges. Only then can a complete 3D model be created.
The new algorithms tackle precisely this critical analysis step. The first algorithm, “Dual 3D Edge Extraction”, recognizes not only sharp but also soft edges in the scan data. This is particularly important as real objects often have both types of edges. The second algorithm, “Opacity-Color Gradation”, creates smooth transitions in both color and transparency. As a result, fine structures are better displayed and the depth perception of 3D edges is improved.
The computing requirements remain within reasonable limits: The researchers were able to successfully test the algorithms on a MacBook Pro with an M2 Max chip and 96 GB RAM. Data sets with up to 108 3D points could also be processed on less powerful systems – for example with Intel Core i7 and 8 GB RAM. The calculation time is at a similar level to conventional methods.
The practical tests show significant improvements: With complex scans, significantly more details and edges are correctly recognized than with previous methods. This could significantly increase the quality of 3D models for 3D printing.
Subscribe to our Newsletter
3DPresso is a weekly newsletter that links to the most exciting global stories from the 3D printing and additive manufacturing industry.
