Abstract:
By analyzing existing disadvantages for spatial object matching methods, multiweighted operator matching method for areal feature geometric matching algorithm is proposed in this paper. This method can reflect four main geometric measuring targets: distance, size, orientation and shape. This proposed algorithm builds six quantitative operators: centroid distance, Hausdorff distance, area overlap, symmetric difference, outline compactness and tangential space. Then we calculate the total similarity for quantitative results by weighted scores. The experimentation used multi-version digital line graphs for matching, and test results show that this method is credible. Furthermore, it can also filter better matching results to more concerned targets by adjusting weight and threshold of the operators.