COORDINATE TRANSFORMATION WITH NEURAL NETWORKS AND WITH POLYNOMIALS IN HUNGARY Ms
暂无分享,去创建一个
Nowadays in Hungary more than one coordinate systems are in use. Frequently it is necessary to transform from one coordinate system to another. The most widely used methods are Helmert and affine transformations. With these traditional methods it is impossible to make a uniform transformation for the whole area of the country, because the attainable accuracy is not sufficient. With polynomial transformation better results can be obtained. The artificial neural network provides a new technology for coordinate transformation. The popularity of this methodology is rapidly growing. In this research coordinate transformations with neural networks were examined between WGS-84 (coordinate system of GPS) and EOV (the Hungarian national reference system), the two most frequently used coordinate systems in Hungary. For calculating the transformation formula one needs to have points with known coordinates in both systems. Two different databases of points were examined. At the beginning of the research a database of 43 points was available. The research was begun with these data, but the number of points was not sufficient for the coordinate transformation with neural networks. Later it a bigger database of 1153 points became available and the examinations were repeated. In both cases (with the two different databases) the transformation with neural network was compared with polynomial transformation.
[1] Ah Chung Tsoi,et al. Universal Approximation Using Feedforward Neural Networks: A Survey of Some Existing Methods, and Some New Results , 1998, Neural Networks.