加权距离排序的水汽层析算法

Water vapor tomography algorithm based on the weighted distance scheme

针对代数重构算法中投影顺序对迭代收敛速度及精度的影响,该文提出了基于加权距离排序的算法。该算法可以优化新选择的投影与先前所用的投影序列之间的角距离,同时能够使得投影分布均匀,从而减小相邻投影之间的相关性;同时对层析方程组按照精度高低进行分组,保证了最终解更接近于实际观测数据;并在所有的实验中使用了非负约束,使得结果符合水汽密度的非负特征。基于香港卫星定位参考站网数据的实验结果表明:加权距离排序算法能够提高代数重构算法的收敛速度与精度。

According to the fact that the order in which the projections are applied in the algebraic reconstruction techniques has a great effect on the speed of convergence and accuracy,the weighted distance scheme was proposed.It heuristically optimized the angular distance of a newly selected projection with respect to an extended sequence of previously applied projections.Also,the selection algorithm assured a uniform sampling of the projection,which can minimize correlation in the projection sequence;the grouping of the chromatographic equations according to the accuracy level ensured that the final solution was closer to the actual observed data;the nonnegativity restriction was added to all experiments,which is consistent with the nonnegativity characteristics of water vapor.Using data from the Hong Kong satellite positioning reference station network,the results indicated the weighted distance scheme had a significant impact on the speed of convergence and accuracy.