earticle

영어

In this study, two coordinate measurement algorithms are described, which are based on one of the multidimensional function optimization methods, so called downhill simplex. One is for calculating relative height of a building, and the other is for calculating the 3D coordinate of a point. We also verify the algorithms mathematically and validate the algorithms by experiments. Two different image pairs are used for the experiments, i.e., IKONOS and Quickbird images, which have been acquired in 2001, 2005, respectively. And the results are compared with the data from GPS survey to evaluate the accuracy. It is shown that the accuracies of the suggested algorithms are within the intrinsic error boundary of the IKONOS and Quickbird images, and they are same as the results obtained by using rigorous sensor model, i.e., differential equations. Further researches are required to apply this method for calculating coordinates from multiple images with heterogeneous sensor models.