GPS坐标转换的三维分离回归模型参数的稳健估计

Three-dimensional Separation Regression the Model Parameters Robust Estimation Method of Transforming GPS Coordinate

为了克服GPS坐标转换为地区坐标传统方法的两个缺陷(坐标数据有粗差,极易出现病态矩阵),致使随机模型中采用了一个不合理的权阵,导致所建数学模型和客观实际存在较大偏差,从而影响最终的转换精度,采用将三维坐标分开转换的策略,在分开转换的过程中应用了抗差估计理论。三维分离回归抗差估计法的实际应用显示该方法能够克服转换中的粗差、病态矩阵的问题。计算数据表明该方法比常规方法在精度上有所提高。

In order to overcome the two primary shortcomings of traditional methods （coordinates data has gross error, another is extremely easy to appear ill-conditioned matrix） for transforming the GPS coordinates into the area sits coordinates, in the cause the weight matrix of the stochastic model are not rational, causes to construct the mathematical model and the objective reality existence big deviation, thus affected the final transformation precision, this article uses the strategy that transform the three dimensional coordinates separately, in the transformation process the robust biased estimators has being applied. The practical application demonstrated this method can overcome gross error, the ill-conditioned matrix question in transformation. The estimated data indicated that this method has the enhancement in the precision than the conventional method.