Point Cloud Reconstruction Based On Intelligent Optimization

Rapid Prototyping Manufacturing Technology (RPM) is a computer-aided design andcomputer-aided manufacturing (CAD/CAM), a successful application in the manufacturingand processing industries. RPM without going through the mold design processes atsignificantly lower production costs, a revolution known as manufacturing. With thedevelopment of society, the people put forward higher requirements of the accuracy of theproduct shape. But due to sweep the surface of the device measurement data error and theproduct itself, constructed product model by PRM is not able to meet the high accuracyrequirements. Surface reconstruction of scattered point cloud as the core of the RPM playingan irreplaceable role to improve the accuracy. As the standard of the manufacturing andprocessing industries in the practical sense, parametric curves and surfaces are efficiency todescribe complex design. So the parametric curves and surfaces generally used as a target inpoint cloud reconstruction. In order to meet the requirement of the accuracy of the point cloudreconstruction technique, the main contents are summarized as follows:We proposed a point cloud reconstruction algorithm based on immune genetic (IGA), ahighly accurate point cloud reconstruction algorithm. In this method, first, the input of thepoint cloud is divided quadrangular patch point cloud. For each patch of point cloud, we firstused the method of least squares B-spline surface approximation and the node vector adaptiveselection of optimal immune genetic algorithm to obtain a higher approximation precision.Finally, numerical approximation algorithms approximate G1continuity of suturing thereconstructed B-Spline patches and then get the final approximation faces.Since the input point cloud has the form of the quadrilateral patches division instead of awhole in the least squares B-spline approximation. The reconstruction algorithm can be moreaccurately describe the complex point cloud model and closed to avoid a complex point cloudparameterization process parameters such as dislocation. In fact, by the slice point cloudreconstruction, it is easily to determine the values of the parameters of the point based on thepoint cloud space distribution. The immune genetic algorithm is the concentration adjustmentmechanism of the immune system introduced into the genetic algorithm. Since the concentration adjustment mechanism of the immune system can effectively inhibit the geneticpopulation with a higher proportion of the number of similar individuals, at the same time beable to promote a lower proportion of the number of similar individuals. Meanwhile it is ableto maintain higher diversity of the population than traditional genetic algorithm and obtainfaster and better convergence results.For optimizing the node vector of the patch has an own best values, and these values aregenerally not identical. If the node vector insertion method to stitch bound may lead to causethe explosive growth of the node vector. Therefore, the numerical approximation method isused to suture the quadrilateral patches. First adjust the boundary control points, then find theoptimum position of the corner control point to obtain the entire patches approximatecontinuous with the help of particle swarm optimization algorithm. At the same time, thepresent algorithm also has the ability of the adaptive approximation secondary breakdown, i.e.If the current division did not meet the accuracy requirements, then the patches again bedivided. thereby obtaining a higher approximation precision.The laboratory results show that the proposed algorithm is enough to obtain the higherB-spline surface approximation precision for complex point cloud model.