Research on enhanced turbine blades point cloud registration based on design benchmarks

Abstract To address the issue of the Iterative Closest Point (ICP) algorithm often yielding optimal registration results instead of accurate ones during the point cloud registration of turbine blades, this study proposes an improved ICP registration algorithm based on design standards. First, the design standard is introduced into the coarse-registered manufacturing point cloud using a ball-filtering algorithm. Then, the Grey Wolf Optimizer (GWO) algorithm is used to solve the registration process and obtain the registration matrix. Finally, the rigidity of the registration is assessed by calculating extrema to determine if it is a local optimum. Iterative adjustments are made until the optimal solution is achieved. Through multiple experiments, it has been shown that the proposed algorithm reduces the average error by approximately 42.9% compared to the classical ICP algorithm and shortens the registration time by about 47.82%.