Abstract:

Aiming at the problems of point pattern matching with large ammounts of non-corresponding points，a truncated secondary spectral matching method based on the K-neighboring local structure is proposed. Truncated spectrum matching is used firstly to get the preliminary result, and the K neighboring local structure is constructed by matching point pairs. The spectrum method is used again in the local structure to find the rest of the matching points. The matching points are merged as the final result. Simulation results show that the proposed algorithm has a good performance in different conditions.