Affiliations: Department of Naval Architecture and Ocean Engineering, Dalian University of Technology, Dalian, China
Note: [] Corresponding author: Guan Guan, Department of Naval Architecture and Ocean Engineering, Dalian University of Technology, Dalian 116024, China. E-mail: guanguan3145@163.com
Abstract: The optimal matching of laser measurements in shipyard with the 3D design model for hull blocks is the key to accurately analyze construction errors and rationally offer modification. The conventional registration algorithms tend to neglect the shipbuilding constraints thus cannot meet the practical shipbuilding demands. To solve the problem, this article presents an advanced matching approach under the actual shipbuilding constraints. Firstly, the initial registration values were obtained by the singular value decomposition (SVD) method. Secondly, according to the accuracy requirement, the error distribution at different directions was determined by weight vectors. Lastly, on the basis of initial registration, an accurate registration that adopted nonlinear multi-objective optimization led to perfect matching by introducing the constraints (e.g. levelness, verticality, flatness and straightness) into the objective function of the computing model. The experimental results proved that the method can obtain the most reasonable matching result satisfying the shipbuilding constraints. It can provide an instructive basis for subsequent assembly.
