Coordinate transformation parameters in Nepal by using neural network and SVD methods

• Main Authors: Ansari Kutubuddin, Gyawali Prabin, Pradhan Prachand Man, Park Kwan-Dong
• Format: Article
• Language: English
• Published: Sciendo 2019-08-01
• Series: Journal of Geodetic Science
• Subjects: coordinate transformation; everest datum; neural network; svd
• Online Access: https://doi.org/10.1515/jogs-2019-0003
• ISSN: 2081-9943

The present study computes B-W extension model (extended Bursa-Wolf model) coordinate transformation parameters from World Geodetic System 1984 (WGS-84) to the Everest datum namely Everest (1830) and Everest (1956) using records of coordinate measurements from Global Positioning System (GPS) observable across Nepal region. Synthetic or modeled coordinates were determined by using the Artificial Neural Network (ANN) and Singular Value Decomposition (SVD) methods. We studied 9-transformation parameters with the help of the ANN technique and validated the outcomes with the SVD method. The comparative analysis of the ANN, as well as SVD methods, was done with the observed output following one way ANOVA test. The analysis showed that the null hypothesis for both datums were acceptable and suggesting all models statistically significantly equivalent to each other. The outcomes from this study would complement a relatively better understanding of the techniques for coordinate transformation and precise coordinate assignment while assimilating data sets from different resources.