利用蚁群遗传算法改进高程异常拟合模型

Improved elevation anomaly fitting model based on ant colony genetic algorithm

针对多面函数在拟合高程异常中难以选取中心节点及光滑因子的问题,该文提出了蚁群-遗传算法改进高程异常拟合模型的方法。为拟合模型构建提供可靠的参数,使选择的中心节点更加合理,加入蚁群算法能够快速获取地形复杂区域的特征点。光滑因子是多面函数拟合法的重要参数,参数值影响了拟合模型精度的高低。采用了遗传算法优化光滑因子,将光滑因子作为种群的染色体进行遗传运算,求得了拟合模型的光滑因子最优值为0.452。利用蚁群-遗传算法改进后的多面函数构建的拟合模型精度为8.6mm,比传统多面函数法拟合结果精度提高了48%。实验研究表明,蚁群-遗传算法改进的多面函数在很大程度上提高了拟合模型的精度,充分验证了改进方法有效可行,为特殊地形的高程拟合提供了重要的参考依据。

In order to solve the problem that it is difficult to select the central node and the smoothing factor in the fitting of the elevation anomaly in multi-face function,a method of improving the elevation anomaly fitting model by ant colony-genetic algorithm was proposed.It provided reliable parameters to the fitting model,which makes the selected center nodes more reasonable,and the ant colony algorithm can quickly obtain the feature points of the complex terrain.Smoothing factor is an important parameter of fitting function of polyhedral function,and the parameter value affects the accuracy of fitting model.The genetic algorithm was used to optimize the smoothness factor,and the smooth factor was used as the chromosome of the population to carry out the genetic operation.The smoothing factor optimum value of the fitting model was 0.452.The accuracy of fitting model built by improved ant colony genetic algorithm was 8.6 mm,and compared with the traditional multi-faceted function method fitting accuracy increased by 48%.Experimental results show that the improved multi-faceted function of the ant colony-genetic algorithm greatly improved the accuracy of the fitting model and fully validated the feasibility of the new method,and provided an important reference for the elevation fitting of special terrain.