Abstract:The high accuracy matching technology of free-form surface was studied. The two-step surface matching method, i.e. the rough matching step and the accurate matching step, was used to obtain higher accuracy results. In the rough matching step, the differential evolution method, which could avoid the local convergence of iterative closest point (ICP) method and overcome the drawback of low convergence rate of genetic algorithm, was put forward to obtain initial points, which would be used in the next accurate matching step. Based on the least square criterion and the least condition principle, the simplex method and the coordinate alternation method were presented to obtain the best matching result, respectively. To improve the behavior of the mathematical model, a scale transform was done with the design variables. With the development tool Visual C++6.0, a computer program was implemented to realize the high accuracy matching of free-form surface. An experiment was conducted to verify the validity and feasibility of these algorithms.